10 1007@b94239

Mol.
Sieves (2004) 4: 337 426

DOI 10.1007/b94239
UV/VIS Spectroscopy
H. Frster
Institute of Physical Chemistry, University of Hamburg, Bundesstrae 45, 20146 Hamburg,
Germany
E-mail: foerster@chemie.uni-hamburg.de
1 Introduction. Electronic Spectroscopy: Ranges, Limits,

Comparison with Related Methods . . . . . . . . . . . . . . . . . . 338
2 Theoretical Fundamentals and Principles . . . . . . . . . . . . . . 339

2.1 Physical Fundamentals and Basic Concepts of Electronic
Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
2.2 Selection Rules, Intensity . . . . . . . . . . . . . . . . . . . . . . . 340
2.3 Fine Structure of Electronic Bands . . . . . . . . . . . . . . . . . . 343
2.4 Types of Electronic Transitions . . . . . . . . . . . . . . . . . . . . 345
2.4.1 Transitions of s, p and n Electrons . . . . . . . . . . . . . . . . . . 345
2.4.2 d-d Transitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
2.4.2.1 Electrostatic Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
2.4.2.2 Molecular Orbital Model . . . . . . . . . . . . . . . . . . . . . . . . 354
2.4.3 Charge-Transfer Transitions . . . . . . . . . . . . . . . . . . . . . . 355
2.4.4 Electronic Transitions in Solids . . . . . . . . . . . . . . . . . . . . 357
2.5 UV/VIS Spectral Changes upon Interaction with the Surface . . . 361
2.5.1 Band Shift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
2.5.2 Change of Absorptivity . . . . . . . . . . . . . . . . . . . . . . . . 362
2.5.3 Change of Bandshape . . . . . . . . . . . . . . . . . . . . . . . . . 364
2.6 Reasons for the Rare Application of Electronic
2.5.4 Appearance of New Bands . . . . . . . . . . . . . . . . . . . . . . . 364
2.6 Reasons for the Scarce Application of Electronic Spectroscopy . . 364
3 Experimental Techniques . . . . . . . . . . . . . . . . . . . . . . . 365

3.1 Absorption/Reflection Spectroscopy . . . . . . . . . . . . . . . . . 365
3.1.1 Dispersive vs. Nondispersive Spectrometry . . . . . . . . . . . . . 365
3.1.2 Comparison Absorption vs. Reflection . . . . . . . . . . . . . . . . 367
3.2 Photoacoustic Spectroscopy . . . . . . . . . . . . . . . . . . . . . . 369
3.3 Luminescence Spectroscopy . . . . . . . . . . . . . . . . . . . . . . 372
4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
4.1 Characterization of the Host . . . . . . . . . . . . . . . . . . . . . . 373
4.1.1 Optical Properties of Zeolites . . . . . . . . . . . . . . . . . . . . . 373
4.1.2 Siting, Oxidation State, Coordination Sphere of Framework
and Nonframework Cations . . . . . . . . . . . . . . . . . . . . . . 375
Springer-Verlag Berlin Heidelberg 2004

338 H. Frster
4.1.2.1 General Aspects of the Influence of the Environment on the

Electronic and Optical Properties of Transition Metal Ions in
Zeolites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
4.1.2.2 Framework Cations . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
4.1.2.3 Nonframework Cations . . . . . . . . . . . . . . . . . . . . . . . . 386
4.1.3 Acid Strength of Zeolites . . . . . . . . . . . . . . . . . . . . . . . 393
4.1.4 Polarity of Zeolites . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
4.2 Characterization of the Guest . . . . . . . . . . . . . . . . . . . . . 397
4.2.1 Inorganic Compounds . . . . . . . . . . . . . . . . . . . . . . . . . 397
4.2.2 Organic Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . 398
4.2.3 Reactive Intermediates: Carbocations, Carbanions, Radicals . . . . 401
4.2.4 Polymerization in Zeolite Channels . . . . . . . . . . . . . . . . . . 406
4.2.5 Clusters of Metals, Metal Ions and Semiconductors
in Zeolite Hosts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
4.2.6 Zeolite-Hosted Oxides . . . . . . . . . . . . . . . . . . . . . . . . . 414
4.2.7 Zeolite-Encapsulated Carbonyl, Metal Chelate Complexes
and Dyes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
4.3 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . 420
5 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
1
Introduction
Electronic Spectroscopy: Ranges, Limits, Comparison with Related Methods
In this chapter, ultraviolet(UV)-visible(VIS) spectroscopy of molecular sieves

will be covered. This constitutes a major part of electronic spectroscopy but has
to be discriminated from other techniques also summarized under this term like
photoelectron, electron energy loss and X-ray absorption spectroscopy. In opti-
cal electronic spectroscopy or electronic spectroscopy in a narrow sense, the
probing particle is a photon that excites an electronic transition without modi-
fication of the incoming radiation. Complementary methods are either emission,
bringing the system back to its ground state, or reflectance, being a combination
of both absorption and emission. Emission or luminescence is divided into flu-
orescence and phosphorescence, respectively, depending on being restricted to or
exceeding the excitation period. In the case of coupling the incident radiation to
an internal process, the outcoming photon has a different frequency.
The energy change of a quantized system is usually described in terms of:
DE = h n = h c/l = h c n (1)
where n is the frequency of the electromagnetic radiation, c is the speed of light

in the vacuum, l is the vacuum wavelength which is obtained by
l = n lair and n c/cair = index of refraction (2)

UV/VIS Spectroscopy 339
Although in the UV-VIS region the radiation is more traditionally characterized

by its wavelength l in units of nm, it is more desirable to use its reciprocal, the
wavenumber n in the units cm1, which has the advantage of being proportional
to the energy. Wavenumbers are always reported in vacuum wavenumbers.
The UV and VIS ranges are somewhat arbitrarily divided and overlap each
other. It seems practical to take the actual or near UV range from 200400 nm or
50,00025,000 cm1 and the VIS range from 400800 nm or 25,00012,500 cm1.
Both are accessible by routine techniques, i.e., by commercial spectrometers. The
neighboring spectral ranges are on the low-energy side at l>800 nm or
n < 12500 cm1 in which essentially vibrational overtones and combination
modes as well as low-frequency electronic transitions are excited, and on the
high-energy side at l<200 nm or n > 50000 cm1, the far or vacuum ultraviolet
passing into the X-ray region where electronic transitions are excited from a core
level into unoccupied bound states or continuum states. This then falls within the
field of X-ray absorption spectroscopy (EXAFS, XANES) and is dealt with in a
separate chapter (see Chap. 5, this Vol.).
In the frequency range 50,00012,500 cm1 information about the properties
of the zeolite host can be obtained, e.g.:
the band gap from transitions between the top of the valence band and the
bottom of the conduction band,
defects like electron-hole pairs,
electronic transitions within the d orbitals of transition metal ions,
charge transfer (CT) processes between cation and anion like transfer of elec-
tron density from a filled oxygen orbital to a partially occupied Mn+ orbital or
intervalence charge transfer (IVCT), i.e., movement of electron density
between metal ions in different oxidation states.
On the other hand, information on the properties of the guest phase can be
obtained provided:
this phase contains chromophores giving rise to, e.g., np*or pp* transi-
tions, or
transitions within the d orbital manifold of transition metal ions, or
charge-transfer transitions, or
transitions between electronic levels explained by the band theory of solids.
2
Theoretical Fundamentals and Principles
2.1
Physical Fundamentals and Basic Concepts of Electronic Spectroscopy
In this section only a breviary of the theory of UV-VIS spectroscopy will be given.
For more detailed information the reader is referred to a selection of textbooks,
e.g., [17].
In the case of isolated atoms purely electronic transitions can be obtained,
while, in the more general case of molecules, rotational and vibrational motions
340 H. Frster
are also simultaneously excited, as illustrated in Fig. 1 by the electronic ground

() and excited states () with their vibrational (v) and rotational levels (J).As the
energies for excitations roughly follow the order DEe>DEv>>DEr103>102 o1, the
position of a rotation-vibration-electronic transition is basically determined by
the change in the electronic energy, while the respective changes of the other
states only give rise to transitions nearby and determine the rotational and vibra-
tional structures of the electronic bands. In the case of polyatomics, and espe-
cially with condensed matter and systems with strong interactions, the rotational
fine structures cannot be resolved at all even with spectrometers of high resolu-
tion, and also a strong perturbation of the vibrational coarse structure has to be
taken into account so that for UV-VIS spectra of zeolitic systems broad unstruc-
tured bands as the envelope are typical. However, in some rare cases, the vibra-
tional structure of electronic bands indicates the superposition of vibrational on
the electronic states.
The intensity of these bands are related to the concentration C and the path
length l by the Bouguer-Lambert-Beer law:
A log (I0/I) = e C l (3)
where A is the absorbance or extinction, I0 and I are the intensities of the mono-
chromatic light before and after the sample, respectively, and e is the extinction
coefficient.
2.2
Selection Rules, Intensity
The UV-VIS spectrum, usually an A or log A vs. n plot, in a first approximation

reflects the discrete electronic states as absorption maxima at different nmax posi-
tions which are correlated with the molecular structure and geometry. The
extinction coefficient e or more likely the integral absorption e (n) dn, which is
approximately the product of emax times the halfwidth Dn1/2, gives information on
the transition dipole moment Rnm or the Einstein coefficient of absorption or
induced emission Bnm which are interrelated by
Bnm = (8 p3/4 p e0 3 h2) |Rnm|2 e (n) (4)

e0 = vacuum permittivity
where e0 is the vacuum permittivity.

The transition moment reflects the probability of a transition between two
states yn and ym as the integral over the volume elements in the coordinates of
all nuclei and electrons dt
Rnm = y*n m ym dt (5)
m is the electric dipole moment operator with components Seixi, Seiyi and Seizi
where ei are the charges on the particles of coordinates xi, yi, zi.
A
Fig. 1. Schematic representation of the vibrational and rotational sublevels of two electronic
states of a molecule
342 H. Frster
The freedom of a possible transition is restricted by the selection rules, a kind of

traffic regulations for spectroscopic transitions, which are based on Eq. (5). The
transition moment being zero or nonzero predicts whether a transition is forbid-
den or allowed, i.e., that the intensity of a forbidden band is much lower in magni-
tude than that of an allowed band. In the case of an allowed transition, the integral
must not vanish, i.e., the integrand must be an even or gerade function, or, in terms
of symmetry, must be or contain the totally symmetric irreducible representation.
The first selection rule concerns the multiplicity of states: Since the compo-
nents of m are ungerade or odd, the integrand becomes gerade or even only pro-
vided the product of the two spin functions is ungerade, i.e., if their multiplicity
or the total spin quantum number S does not change, or in other words if DS=0.
Thus singlet-triplet transitions are normally forbidden.
The second selection rule arises from the symmetry of the two states: Since the
components of m transform like the translations, the direct product of the inter-
connected wave functions must also transform like at least one of the translations
in order to give the totally symmetric representation.
The third selection rule is valid for molecules with a center of symmetry: Only
transitions changing the parity, i.e., where gu, are allowed (Laporte rule).
Although forbidden, such transitions are frequently observed as bands of
finite intensity:
in the case of many-electron excitation due to configurational interaction
due to intramolecular or intermolecular perturbations by interaction with the
surface of an adsorbent or by distortion of the molecular symmetry from
simultaneously excited vibrations.
In order to quantify the intensity of an electronic band the oscillator strength f
is used. This dimensionless quantity relates the observed integrated intensity to
that of a calculated integrated intensity according to a simple model where the
excited electron is attracted to the center of the molecule with a Hookes law type
force, i.e., using harmonic oscillator wave functions. f is the ratio of observed and
calculated integrated intensities and amounts to
f = e(n) dnobs/ e (n) dncalc = (4 e0 m c2 ln 10/NL e2) e (n) dnobs (6)
where m is the mass of the electron and NL the Loschmidt number and should
give a maximum value of 1. If fO1 the transition is forbidden.
According to Eq. (1), electrons can be excited from the ground state S0, which
is usually a singlet (S) state, into higher electronic states S1, S2 (Fig. 2). The
return to S0 can occur:
by fluorescence, i.e., emission of radiation,
by non-radiative transitions, i.e., internal conversion (IC),
by consecutive intersystem crossing (ISC) steps under spin flip from the sin-
glet to the triplet and back to the singlet states: S1T1S0, or
by phosphorescence which is a spin-forbidden radiative T1S0 intercombi-
nation transition.
As transitions are possible between electronic, rotational and vibrational
states, the system in a given electronic state can relax to lower rovibrational
Fig. 2. Term diagram for electronic transitions. A Absorption; F fluorescence; Ph phosphores-

cence; IC internal conversion; ISC intersystem crossing. Reprinted from Hesse M, Meier H, Zeeh
B (1995) Spektroskopische Methoden in der organischen Chemie, with kind permission of
Georg Thieme Verlag, Stuttgart
states, so that the fluorescence transition is red-shifted compared to the excita-

tion band.
2.3
Fine Structure of Electronic Bands
A change in the electronic state by DEe would give rise to absorption or emission
of one line. The simultaneous change DEv of the vibrational states should result
in a progression of so-called vibronic lines while the additional change DEr of the
rotational energies transforms it to a progression of rovibronic bands, which is
a series of transitions, in the case of absorption with a common lower (v), in the
case of fluorescence or phosphorescence with a common upper (v) level. Figure 3
depicts the superposition of electronic with vibrational (and rotational) excita-
tion for the case of internuclear distances in the two electronic states not signif-
icantly different from each other. As according to the Boltzmann expression
Nv/Nv=0 = exp [(Ev Ev=0]/kT] (7)
higher vibrational levels v= 1, 2, are also appreciably populated, additional

and partly overlapping progressions arise from these hot v states.
Due to the fairly large energies the electronic transitions are accompanied by
changes in the electronic structure reflected in alterations of the potential energy
curves of the excited compared to the ground state (Fig. 4). For the former the
equilibrium distances usually have increased as a result of bond weakening due
to the electronic redistribution. On the changes of the vibrational states during
electronic excitation there are no general restrictions, e.g., selection rules. The
intensity distribution of the vibrational structure of an electronic band can be
explained by the Franck-Condon principle: The electron jump takes place so
rapidly in comparison to the vibrational motion that immediately afterwards the
nuclei still have nearly the same relative position and velocity as before, i.e., the
344 H. Frster
Fig. 3. Electronic excitation between singlet states superimposed by vibration giving rise to a
vibrational progression for absorption (A), fluorescence (F) and phosphorescence (Ph).
Reprinted from [7] with permission of Springer, Berlin Heidelberg New York. Copyright 1992
nuclei may be regarded as fixed during the transition. Taking into account
the probability of being at a particular internuclear distance an electronic tran-
sition is most favored at a position such that it connects probable states of
the system. In emission the most intense lines can be understood by vertical
transitions starting from internuclear distances near the turning points of the
oscillator. From quantum mechanical treatment the transition moment Rev is
given by
Rev = y*e me ye dte yv yv dtv = Re yv yv dtv (8)
where me is the part of the dipole moment depending on the electrons, te and tv
are the electronic and vibrational coordinates, respectively, and Re is the elec-
tronic transition moment which for a given electronic transition is a constant.
Thus, the second integral, the vibrational overlap integral, a measure of the over-
lap of the two corresponding vibrational wave functions, becomes important.
Figure 4 shows the intensity distributions of vibrational progressions with
increasing differences between the equilibrium nuclear distances in the elec-
tronically excited and ground state.
Fig. 4. Franck-Condon principle illustrated for the case re>re and typical intensity distribu-
tions of vibrational progressions for the three cases indicated below. Reproduced from [6] with
permission. Copyright John Wiley and Sons, Ltd
2.4
Types of Electronic Transitions
2.4.1
Transitions of s , p and n Electrons
Analogous to the concept of group vibrations in IR spectroscopy, photon absorp-

tion in the UV-VIS range can be frequently traced back to the excitation of spe-
cific electrons in a particular group of the system, called a chromophore, that is
responsible for the color of the respective compound. Substituents that on their
own do not confer color but increase (hyperchromic effect) or decrease
(hypochromic effect) the coloring power of a chromophore are designated as
auxochromes or antiauxochromes. Groups that shift the absorption maximum
346 H. Frster
to lower or higher wavenumbers are called bathochromic or hypsochromic

groups.
Electrons in polyatomic systems, essentially s and p electrons, may occupy
bonding, antibonding or nonbonding molecular orbitals (MOs). According to
MO notation, bonding electrons are distinguished into s or p electrons if their
wave functions or charge densities are rotationally symmetric with respect to or
contain a nodal plane through the valence axis. Electrons in the corresponding
antibonding orbitals are labelled with asterisks and denoted s* or p* electrons.
Unshared or nonbonding electrons are called n electrons. The sequence of
increasing electronic energy is usually in the order: s<p<n<s*<p*<Rydberg
(R) states. The latter are highly excited states near the ionization limit.
To date there is no accepted standard notation for electronic transitions.With-
out going into detail some of the frequently used notations will be discussed
shortly.
The first notation, introduced by Kasha, stems from MO representation. Nor-
mally, electrons are raised from bonding or nonbonding orbitals to empty anti-
bonding orbitals. As transitions between p and s orbitals are excluded by sym-
metry, in principle the following types remain: ss*, pp*, np*, and n s*.
Both the arrows and are used at will so that the meanings of pp* and
p*p are the same and, as the direction is indicated, there will be no confusion
about the ground and the excited states; np* transitions are also symmetry-
forbidden, resulting in bands of very low intensity. The ranges of absorption for
the above-mentioned transitions are outlined in Fig. 5.
The notations after Mulliken use different letters for denoting the ground and
excited states, as shown in Table 1.
Fig. 5. Ranges of absorption for different electronic transitions. Reprinted from Hesse M, Meier
H, Zeeh B (1995) Spektroskopische Methoden in der organischen Chemie, with kind permis-
sion of Georg Thieme Verlag, Stuttgart
Table 1. Different notations of electronic transitions
System Symbols Symbols Examples Correspondence

Ground state Excited state
Kasha s, p, n s*, p* ss*

(MO pp*
representation) np*
ns*
Mulliken N Q, V, R VN pp* or ss*
QN np* or ns*
RN Rydberg transition
Group theory Irred. repr. Irred. repr. 1A 1A
1 1
incl. incl. 1B 1A
2 1
multiplicity multiplicity 1E 1A
1u 1g
Finally, the group theory notation classifies the states by the symmetry behav-
ior of their electronic wave functions.
Let us consider ethylene as an example. From the twelve valence electrons ten
are s electrons while two occupy a bonding p MO, yb=2ppA+2ppB=pu between
the carbon atoms A and B. The corresponding antibonding combination is a
ya=2ppA2ppB=pg* MO. According to the D2h point group these two orbitals
belong to the b3u and the b2g* irreducible representations. The ground state con-
figuration is yb2 =b3u2 with paired spins due to the Pauli exclusion principle, i.e.,
forming a singlet. Configuration and state are characterized by the symmetry
species of the corresponding point group using for distinction lower-case letters
for the former and capital letters for the latter. Thus, the ground state is totally
symmetric and, including multiplicity 2S+1, denoted by 1Ag. Exciting one p elec-
tron into the antibonding MO leads to the configuration ybya=b3ub2g* with either
paired (singlet) or unpaired spins (triplet), the direct product of which gives the
symmetry species 1B1u or 3B1u, respectively. According to Hunds rule the triplet
state has the lower energy. Raising both electrons into the pg* MO would give rise
to the configuration ya2 =b*2g2 and thus again to a state 1Ag. From the ground and
the three excited states three transitions should be expected, which in the differ-
ent notations are shown in Table 2. The first is forbidden as a result of intersys-
tem crossing. The second transition is allowed as the direct product of the inter-
connected wave functions. 1Ag1B1u=1B1u belongs to the same representation as
the z coordinate (and therefore is polarized in the z direction), resulting in a non-
Table 2. The three lowest electronic transitions of ethylene in different notations
Configuration Kasha Mulliken Group theory Selection rule
yb yayb2 pp* TN 3B , 1A
1u g forbidden: DS0
yb yayb2 pp* VN 1B 1A
1u g allowed: 1Ag1B1u=1B1u
ya2yb2 pp* VN 1A 1A
g g forbidden: gg
348 H. Frster
vanishing transition moment Rnm and thus in a strong band near 165 nm. The
third 1Ag1Ag transition is parity-forbidden and only gives rise to a weak band
around 210 nm.
According to the building-up principle in molecules, the MOs are occupied in
a distinct order with increasing energy. The two so-called frontier orbitals, the
highest occupied molecular orbital (HOMO) with energy En and the lowest unoc-
cupied molecular orbital (LUMO) with energy En+1, play an important role in
explaining the optical properties of these compounds. Upon exposure an electron
is promoted from the HOMO into the LUMO. This so-called free electron mole-
cular orbital model has been applied to molecules with conjugated double bonds
and to aromatic hydrocarbons. On the basis of an electron-in-a-box approach the
N-delocalized p electrons of the double bonds occupy in pairs the n=N/2 energy
levels up to the HOMO. Thus, the frequency of the HOMOLUMO transition can
be derived from:
n = (En+1 En)/h = (2n+1) h/8 m a2 = (N+1) h/8 m a2 (9)
with m the electron mass and a the length of the conjugated system.
2.4.2
dd Transitions
While in the preceding sections the discussion was mainly about s and p elec-
trons, a second type of electronic transitions, typical of transition-metal com-
pounds, concerns the excitation of d (or for rare earth elements f) electrons, espe-
cially in the presence of a ligand or crystal field. In the case of transition metals
the ligand field partly removes the degeneracy of the d orbitals so that dd tran-
sitions become feasible. This phenomenon can be explained either by (i) an elec-
trostatic or (ii) a MO model. Here, only a brief introduction into the general fea-
tures will be given. Further details can be found in the literature [e.g., 812].
2.4.2.1
Electrostatic Model
Here the environment is modelled by point charges. As the five d orbitals of the
central ion are directed either along (dx2y2, dz2) or diagonally between the coor-
dinate axes (dxy, dyz, dxz), the approach of charges in a distinct arrangement with
respect to the coordinates gives rise to a different perturbation of these orbitals
and thus to a splitting caused by the environment. In a field of octahedral sym-
metry, the dx2y2 and dz2 orbitals which belong to the eg representation are raised
by 3/5 Do, while the dxy, dyz, and dxz orbitals of t2g symmetry are pushed down by
2/5 Do, where Do is the energy difference egt2g increasing with the strength of the
crystal field. In order to avoid fractions the definition Do10 Dq has been intro-
duced where, according to Schlapp and Penney [8], Dq represents the field
strength parameter. In this alternative notation, the eg orbitals lie 6 Dq above and
the t2g orbitals 4 Dq below the average. In an octahedral surrounding the dd
transitions from t2g to eg are Laporte-forbidden and the respective bands in the
VIS and NIR regions are weak in intensity.
Fig. 6. d Orbital splitting in ligand fields of different symmetries
In tetrahedral environments the splitting becomes reversed: the former t2g

orbitals, now of t2 symmetry, are lifted by 2/5 Dt and the former eg orbitals, now
of e symmetry, are shifted downwards by 3/5 Dt, where Dt=Et2Ee=4/9 Do (see
Fig. 6). Upon lowering the symmetry of the environment the residual degener-
acy is removed further as shown, e.g., for a square planar field.As there is no cen-
ter of inversion, the Laporte selection rule is no longer valid resulting in an
increased intensity of the electronic bands.
350 H. Frster
So far only the interaction with the external field has been considered. But as
a counterpart the electronic interaction also has to be taken into account. As
known from textbooks, each electron configuration splits due to electrostatic
(i.e., Coulomb) repulsive interaction into states 2S+1L of different energies with
multiplicity 2S+1 that are (2L+1)(2S+1)-fold degenerate, and some of these states
are forbidden by the Pauli exclusion principle. From the remaining states, accord-
ing to Hunds rules, the term with the greatest multiplicity and, for a given mul-
tiplicity, the term with the greatest L value, lies lowest in energy, i.e., forms the
ground state.
A d2 configuration, for example, splits into states of energetic order:
3F(21) < 1D(5) < 3P(9) < 1G(9) < 1S(1)
with degeneracies given in parentheses.

From quantum mechanical treatment of the term energies two types of inte-
grals, the Coulomb integral Jij and the exchange integral Kij, have to be evaluated
[8, 12].Without going into detail, it has turned out to be advantageous to express
these integrals in terms of parameters A, B and C, the so-called Racah parame-
ters, which quantitatively express the meaning of Hunds rules. These parameters
are positive quantities since they represent electron repulsion, the linear combi-
nation of which makes possible the calculation of the term energies using the fol-
lowing expressions:
E(1F)=A8B, E(3P) = A + 7B,
E(1G)=A+ 4B+ 2C, E(1D) = A3B + 2C (10)
and E(1S)=A+14B+7C.
As A is common to all terms, and if one is merely interested in relative energies,
only the parameters B and C must be known. Further simplification can be
achieved by the approximation C4B.
In the case of an octahedral field (see Fig. 6) the ligand field stabilization
energy (LFSE), which is the net energy of a (t2g)x(eg)y configuration relative to the
average energy of the orbitals, is obtained from:
LFSE = (2/5 x + 3/5 y) Do (11)
On the other hand, an electron which pairs with an electron already present in
an orbital experiences a strong Coulomb repulsion, the so-called electron
pairing energy P. So, the net energy gain upon occupying the orbitals at a given
configuration is:
LFSE + P = (2/5x + 3/5y) Do + P (12)
and, therefore, dependent on Do, i.e., on the ligand field strength. In the weak-field
case, Do<P, occupation of the eg orbital, in the strong-field case, Do>P, electron
pairing is more favorable, giving rise to unambiguous configurations in the cases
d1d3 and d8d10, but to two different high-spin and low-spin configurations in
the cases d4d7, as shown in Fig. 7.
Fig. 7. Electron configurations in octahedral environments. Reproduced from [6] with per-
mission. Copyright John Wiley and Sons, Ltd
For labelling the states of an electronic transition spectral term symbols are
used in the form 2S+1G, where G is the upper-case version of the symmetry species
of the overall orbital state and 2S+1 the multiplicity. For the establishment of the
spectral terms two extreme cases are considered:
the weak-field limit in which only electronic repulsion is important
dn 2S+1L 2S+1G
the strong-field limit in which only the ligand field is determining
dn (t2g)ny (eg)y 2S+1G
both meeting at the same intermediate case 2S+1G, which is obtained by drawing
a correlation diagram (Fig. 8), exemplified for the d2 configuration.
In the weak-field limit due to electron repulsion the d2 configuration splits into
a couple of states 3F<1D<3P<1G<1S, as shown before. The energy of the 3P state
relative to the 3F ground state, expressed by means of the Racah parameters, is
15B according to Eq. (10). Switching on octahedral (Oh) field interaction, due to
group theoretical considerations the 3F state correlates with 3T1+3T2+3A2 and the
3P state with 3T under conservation of the multiplicity.
1
352 H. Frster
Fig. 8. Correlation diagram for d2 ions in octahedral environment. The subscript g has been
omitted as being common for all states and orbitals. Quoted from [9]. Copyright 1971 by John
Wiley & Sons, Inc. Reprinted by permission of John Wiley & Sons, Inc
In the strong-field limit, a d2 ion has the configurations:
(t2g)2 < (t2g)1 (eg)1 < (eg)2
According to Eq. (11), the relative energies of the (t2g)1 (eg)1 and (eg)2 states with
respect to (t2g)2 are Do and 2Do, respectively. The symmetries of the orbital states
produced by electronic repulsion can be determined by taking the direct prod-
uct and decomposing into 3T1g+1T2g+1Eg+1A1g for t2g 2 , into 3T +3T +1T +1T for
2g 1g 2g 1g
1 1 3 1 1
t2g eg and into A2g+ Eg+ A1g for eg .2
A schematic overview of term splitting, starting from the strong-field limit

(right) and from the weak-field limit (left), meeting at the actual case, i.e., taking
into consideration repulsion as well as field interaction, is given in Fig. 8.

As ISC is forbidden and the ground term is a 3T1 state, it is satisfactory to con-
fine it to the triplet term from which transitions to 3T2, 3T1 and 3A2 can be
expected. The energies of the intermediate cases with both ligand field and
electron interaction involved are obtained by correlating both limiting cases
which is now a standard problem in group theory. After evaluation of the sym-
metries of the split states, quantum mechanical calculations provide these
energies as a function of the ligand field strength. The most frequently used
graphs are the Tanabe-Sugano diagrams in which both energies E and the
ligand field strength D are expressed in units of the Racah parameter B, and
E/B is plotted vs. D/B taking the ground term energy of the free ion as zero
(Fig. 9). From the analysis of these diagrams it becomes obvious that:
the multiplicity does not change upon ligand field splitting,
some of the lines are curved due to interaction of terms of the same symme-
try which mix, as they obey the noncrossing rule of quantum mechanics, and
diverge from each other,
the diagrams become discontinuous when the ground term changes due to
preference of electron pairing in strong ligand fields, and
Fig. 9. Tanabe-Sugano diagram for the d2 configuration. Quoted from [9]. Copyright 1971 John
Wiley and Sons, Inc. Reprinted by permission of John Wiley and Sons, Inc
354 H. Frster
the diagrams are in pairs qualitatively identical: those of the dn systems in an

octahedral field resemble those of the d10n systems in a tetrahedral field, and
vice versa, explainable by the particle-hole formalism.
The electron repulsion in a free ion in the gas phase is always larger than that in
a complex and can be expressed by the nephelauxetic parameter b as the ratio of
their Racah parameters:
b = B (complex)/B (free ion) < 1 (13)
The term b depends on the ligand descending in the so-called nephelauxetic

series:
F > H2O > NH3 > CN, Cl > Br > I
in the sequence of increasing covalent character of the bonding between ion and
ligand. Due to the overlap of the metal and ligand orbitals, the space of electron
residence becomes enlarged, resulting in a lower electron-electron repulsion. The
simple electrostatic model has proved inadequate to explain this series and thus
the field splitting D.
2.4.2.2
Molecular Orbital Model
In the case of strong interaction between the central ion and ligands the crystal
field approximation breaks down and for explanation the molecular orbital (MO)
model is preferred. Within its frame the environment is no longer considered as
an arrangement of point charges. Instead, bonding/antibonding molecular
orbital pairs are formed in the usual way from the orbitals of the central ion and
the s and the p orbitals of the ligands provided they are of the same symmetry.
This is schematically shown for transition metals of the fourth period in an octa-
hedral environment neglecting the p electron interaction in Fig. 10. For the t2g
orbitals of the central ion there is no symmetry counterpart among the ligand s
orbitals (a1g, t1u and eg in Oh symmetry), and so they are adopted as nonbonding
orbitals into the MO scheme. Including the ligand p orbitals (t1g, t2g, t1u and t2u
in Oh symmetry) gives rise to changes of the MO levels especially of the t2g
orbitals which, in this case, turn into bonding/antibonding orbitals, whereas now
the ligand t1g and t2u orbitals have become nonbonding MOs due to the absence
of symmetry counterparts among the metal ion orbitals. The MO model is more
versatile in explaining the increase in the field splitting D by the formation of p
bonds with suitable ligand orbitals. However, compared to the electrostatic
model, the clarity of imagination is lost.
The Jahn-Teller theorem states that in nonlinear systems a degenerate ground
electronic configuration is unstable and becomes stabilized by distortion of
the Oh or Td symmetries to lower, e.g., tetragonal or, as an extreme, to square
planar symmetries, respectively, thus removing the degeneracy. In this way the
energy separations between the ground and higher levels are increased and
Fig. 10. MO diagram of octahedral complexes under neglect of p electron interaction
additional transitions become feasible. Typical cases are d4, d7 and d9 complexes
with Eg and T1g ground states. Jahn-Teller distortion is also valid for excited
states.
The intensities of d-d bands are generally low as being forbidden by the par-
ity and sometimes also by the multiplicity selection rules. Although forbidden,
the observed intensities can be explained by relaxation of the selection rules
by distortion, i.e., deviation from perfect centrosymmetry, vibronic interaction,
configuration interaction or spin-orbit coupling, the details of which are beyond
the scope of this article.
2.4.3
Charge-Transfer Transitions
Charge-transfer (CT) transitions take place in systems made up of two compo-

nents, electron donor (D) and acceptor (A), correctly designated as donor-accep-
tor complexes. The binding strength between the two species extends from strong
bonding, as with the transition metal-ligand complexes shown before, up to the
other extreme, transitory contact CT complexes, which exist only during the
356 H. Frster
duration of collisions between donors and acceptors. It is suggested that donor-

acceptor complexes are formed according to:
D + A = DA D+ A (14)
where DA and D+A describe different resonance structures, the interaction of
which is responsible for the resonance energy stabilization of the complex. The
ground state may be expressed by its total wave function:
yo = y (DA) + ly(D+ A) (15)
where l2 determines the ionic contribution to the ground state. Electron ex-
citation is connected with charge rearrangement involving charge transfer from
one component to the other, expressed by the total wave function of the excited
state:
y1 = y(DA)+ ly(D+ A) (16)
and l>l indicates the excited state to be more ionic. Any endeavors to form
a donor-acceptor complex and the energy change in the transition depend on
(i) the ionization potential of the donor and (ii) the electron affinity of the
acceptor.
There are several types of donor-acceptor interactions, the energy of which
increases in the sequence: contact pairs<ps<pp<ns<Lewis acidLewis base.
In the case of transition-metal complexes one has to distinguish between (i)
ligand-to-metal charge-transfer transitions (LMCT) and (ii) metal-to-ligand
charge-transfer transitions (MLCT). For example, with the former, lone pair lig-
and p orbitals not directly involved in bonding are promoted into an empty
orbital of predominant metal character. LMCT transitions require high-lying
lone pair orbitals and low-lying empty metal orbitals, a prerequisite which is met
in the case of high oxidation numbers of the metal equivalent to a low d-orbital
population. MLCT transitions, on the other hand, are observed with metals of low
oxidation numbers and empty low-energy p* ligand orbitals. Examples of this
are the spectra of complexes with aromatic, carbonyl and cyano ligands.
Upon dissolving of salts characteristic bands appear, which have to be attrib-
uted to charge-transfer-to-solvent transitions (CTTS) being dependent on the
solvent polarity. As these are allowed by the selection rules in most cases, the
intensity of CT bands is generally rather strong and depends on the extent of
overlap of the yo and y1 wave functions.
In luminescence experiments donor-acceptor complexes show the mirror
image relationship between absorption and emission spectra. An interesting
anomalous feature has been found with typical donor molecules in solutions of
aromatics: A broad structureless emission band arises at the low-frequency side
of the hydrocarbon band increasing in intensity with growing donor content at
the expense of the former. This phenomenon has been explained by the forma-
tion of CT complexes in the excited state, so-called exciplexes. Exciplexes may be
formed by interaction of excited acceptors with donors (A*+D) or reverse
(D*+A).
2.4.4
Electronic Transitions in Solids
From the viewpoint of the tight-binding approximation, the electronic structure

of a solid is characterized by delocalization of the valence electrons throughout
the entire structure. The overlap of a large number of atomic orbitals (AOs)
results in closely spaced molecular orbitals (MOs), so-called bands, designated as
s-, p-, d-bands and so on, according to the angular momentum of the related
atomic state. Depending on the separation of the AOs and on the strength of their
interaction there may be a gap between them, a range of energies of zero density
of states. This gap may be absent in the case of strong interaction where these
bands overlap. The uppermost band is called the conduction band if its filling is
incomplete, but it is called the valence band if it is full. In this case the conduc-
tion band is the empty band just above. With the band structure model the dif-
ferent properties of metals, semiconductors and insulators have been successfully
explained. As far as radiative transitions are concerned, typical situations for
photon absorption are indicated in Fig. 11: A conductor with a partially filled
conduction band (a) shows excitations to nearby empty states in the same band
or, surmounting the energy gap, to states in the uppermost empty band, reflected
also by a gap in the spectrum. A conductor with superposition of the uppermost
bands (b) thus shows a continuous absorption. In an insulator with a filled
valence band (c) only transitions to the conduction band separated by a large gap
are possible. Decrease of the gap would explain the behavior of semiconductors.
The final example (d) describes the situation in the case of lattice defects, i.e.,
Fig. 11ad. Radiative transitions in solids. Conductor a with incompletely filled conduction
band; b with band overlap; c insulator or semiconductor depending on the width of the gap;
and d insulator with lattice defects. From Alonso M, Finn EJ. Fundamental University Physcs,
vol VIII. Copyright Addison-Wesley Publishing Company. Reprinted by permission
358 H. Frster
imperfections of structure or composition, normally present in all solids as a

consequence of equilibrium at nonzero temperatures predicted by thermo-
dynamics. Intrinsic point defects are either vacancies or framework particles
displaced into interstitial sites. Extrinsic point defects, on the other hand, are
due to impurities. Intentionally introduced impurities are called dopants. Dop-
ing of solids with atoms of an element with more or fewer electrons gives rise to
the formation of narrow donor or acceptor bands in the band gap and is respon-
sible for n-type or p-type semiconductivity. In addition to the band gap crossing,
much lower discrete energies, e.g., around Ei, are required for electron excitation.
Provided these energies fall into the visible or near ultraviolet range, the respec-
tive point defects are called color centers.
The properties of color centers have been most thoroughly investigated
with alkali halides. From the multitude of color centers the most frequently
observed ones are the so-called F-centers (from German Farbzentren) obtained
either (i) by thermal treatment of alkali halides in alkali metal or halogen vapor,
(ii) by irradiation with UV light, X- and g-rays as well as energy-rich protons,
or (iii) by solid-state electrolysis. Only in case (i) is there a deviation from stoi-
chiometry.
The F-center is equivalent to an electron trapped in an anion vacancy and,
being nonlocalized, is extended over a space of more than the lattice constant.
The electron spends more residence time near to all alkali atoms adjacent to the
vacancy, forming temporarily a sodium atom.
Energetically, an F-center is equivalent to a number of discrete states localized
in the gap between the valence and the conduction band. Optical absorptions are
due to electronic transitions from the ground to excited states, the position of
maximum absorption lmax being almost proportional to the square of the lattice
constant.
Besides F-centers a number of further centers can be formed by thermal and
optical treatment:
1. Centers with trapped electrons. The bands starting from the ground state to
excited states are designated in the sequence of increasing energy as F-, K- and
L-bands. In the case of the latter, promotion into the conduction band is pos-
sible. F-centers are formed during bleaching the color caused by irradiation
into the F-band and attributed to anion vacancies occupied by two electrons,
which can also form by trapping of electrons from the conduction band in
F-centers. The F-band corresponds to the ionization of the second weakly
bound electron. More complicated color centers are formed by aggregation.
Their respective absorptions are named M-, R-, N- and O-bands.
2. Centers with trapped holes. For instance, bands due to Cl2 on anion lattice site
absorbing at the high-frequency side of the F-band are called H- and V-bands.
3. Exciton centers. Excitons are mobile, excited electronic states connected with
the formation of an electron-hole pair, equivalent to a positronium atom,
which carry energy but no charge through the solid. Excitons can be consid-
ered as polarization waves spreading out in the crystal and can be described
as lying between two limiting cases distinguished as (i) the Frenkel exciton (in
alkali halides) and (ii) the Wannier-Mott exciton (in semiconductors).
The former is characterized by a radius small compared to the lattice para-

meters, and its energy of formation Ef in alkali halides is well reproduced by
the empirical Hilsch-Pohl formula:
Ef =AI+Me2/a (17)
where A is electron affinity of the halogen, I the ionization potential of the alkali
atom, M the Madelung constant and a the nearest distance between the ions.
The latter is equivalent to an electron in the conduction band and a hole in
the valence band, i.e., with a radius large compared to the lattice parameters.
Therefore, bonding of the electron is weak similar to a positronium atom
embedded in a dielectric with dielectric constant e and an interaction poten-
tial approximated by V(r)=e2/er. Such a system has energy states similar to
the hydrogen atom following the formula:
E =Eg 4 p2 e4/2 h2 e2 n2 (18)
where Eg is the energy gap, m is the reduced mass of an electron-hole pair and
n the principal quantum number. According to Eq. (18), the exciton ground
state with n=1 is Eg4p2e4/2h2e2 above the valence band edge, and the excited
states are bands located within the gap. Exciton transitions occur near the low-
energy side of the band gap.
4. Centers due to impurities.Another type of color centers are impurity ions, the
bands of which are caused by electronic transitions to neighboring ions of the
host lattice (U-centers) or of electron defects trapped by impurities (FZ-cen-
ters). In the case of high impurity concentrations colloidal segregations may
form.
The process of luminescence in solids is also influenced by the band structure,
as schematically outlined in Fig. 12. Figure 12a illustrates the promotion of an
electron into the conduction band leaving behind a hole in the valence band. In
Fig. 12b an electron of one of the impurity levels fills this hole while an electron
of the conduction band may fall into an empty impurity level, which in Fig. 12c
reduces its energy by transition into the impurity ground state, emitting radia-
tion of lower frequency than the incident one. Instead, the electron may be
trapped in a defect from which a radiative transition to the impurity ground state
is forbidden (Fig. 12d). In this case, being a metastabile state, the transition may
be either delayed giving rise to phosphorescence or the electron may be liberated
by excitation via the conduction band (Fig. 12e) and then go the way of normal
fluorescence according to Fig. 12b and 12c.
In the case of metals their UV-VIS properties can be described by quantized
plasma vibrations, so-called plasmons. Plasmons can be excited by inelastic scat-
tering of electrons or reflection of electrons or photons on thin metal layers. More
details are available in the literature [13].
360
Fig. 12ae. Mechanisms of luminescence in solids. From Alonso M, Finn EJ. Fundamental University Physics, vol VIII. Copyright
Addison-Wesley Publishing Company. Reprinted by permission
H. Frster
2.5
UV/VIS Spectral Changes upon Interaction with the Surface
In this section the influence of molecular interaction on solid surfaces in general

will be outlined; this topic has been reviewed in more detail in the literature
[1416]. There is a broad range of interactions reaching from weak to strong,
placed between the extreme cases of physisorption and chemisorption and prob-
ably at best characterized by the magnitude of heat released during adsorption.
Spectroscopy is of special interest as a tool for characterization as it is sensitive
to all changes of the molecular properties due to interaction with the surface as
well as lateral interaction with adjacent molecules. Furthermore, information
about the nature of the active surface sites may be obtained.
UV-VIS spectroscopy in the first place is sensitive to alterations in the elec-
tronic structure. Studies of molecular interactions on solid surfaces started much
later than those in solution. It is of historical interest that already in the early
1930s, UV-VIS spectroscopy was one of the first techniques applied to a zeolite,
as shown in Fig. 13 for iodine adsorption on chabazite [17]. The figure shows
the spectrum of iodine on chabazite compared to those of iodine as vapor, as a
solid, dissolved in benzene and adsorbed on calcium fluoride. As illustrated by
these examples, four main effects of interaction on the spectrum can be estab-
lished [15, 16]:
1. shift of the band position,
2. change of the absorptivity,
Fig. 13. Historical UV-VIS spectra of iodine adsorbed on chabazite and calcium fluoride com-
pared to solid I2, I2 vapor (broken line) and I2 dissolved in benzene. Reprinted from [17] with
kind permission of R. Oldenbourg Verlag GmbH, Mnchen
362 H. Frster
3. change of the bandshape with respect to its vibrational structure, and

4. appearance of new bands.
2.5.1
Band Shift
The surfatochromic shift is characterized by magnitude and direction and is

mostly related to the absorption maximum which, depending on the Franck-
Condon factor, may not necessarily be the 00 but more likely a 0v transition.
An explanation for both effects was given by de Boer [18] who, as shown in
Fig. 14, used the concept of potential energy diagrams of a molecule A in the
ground and the excited state A* when approaching a solid surface together with
the Franck-Condon principle: The energy hng=ad necessary to promote an elec-
tron in the free molecule, i.e., far away from the surface, changes to that for the
vertical transition in the adsorbed state hna=be and can be related to the heats
of adsorption QA=DH in the ground and in the excited state QA* by:
hna h ng = QA QA* (19)
resulting in a frequency shift
Dn = na ng = (QA QA*)/h (20)
which depends on the difference QAQA*. QA* is not completely identical to the
heat of adsorption of the excited state QA*>QA*, but rather a hypothetical heat of
adsorption which should be released provided the excited admolecule remains
at the same distance from the surface as in the ground state.According to Eq. (20)
three different cases must be distinguished:
1. QA>QA* gives rise to a hypsochromic shift,
2. QA<QA* leads to a bathochromic shift, while for
3. QA=QA* the band position is unchanged.
From derivation it becomes clear that the frequency shift is not a measure of the
absolute strength of interaction as it reflects only their differences in the excited
and the ground state.
The two-dimensional treatment does not consider the change in the intramol-
ecular nuclear distance in the two states which has been taken into account as a
three-dimensional representation by de Boer and Custers [19].
2.5.2
Change of Absorptivity
Generally, upon adsorption, the intensity of UV-VIS bands is significantly altered

caused by an increase or a decrease in the extinction coefficient e. This effect
depends mainly on the adsorption geometry, i.e., whether the electronic vector
of the adsorbed species is parallel or perpendicular to the electrostatic surface
field. In addition, e may follow a direct or inverse variation with coverage q, some-
Fig. 14A, B. Potential energy curves of adsorbed molecules in the ground (lower curve) and the
excited state (upper curve) for the explanation of bathochromic (A) and hypsochromic spec-
tral shifts (B) according to de Boer. Reprinted from [18] with kind permission of R. Oldenbourg
Verlag GmbH, Mnchen
times by several orders of magnitude, depending on both the adsorbate and the
adsorbent. In the case of a heterogeneous surface this behavior parallels that of
the dependence of the heat of adsorption on coverage: At low loading the per-
turbation of the admolecule is greatest due to interaction with the most active
surface sites and then decreases when these sites are occupied. The resultant
extinction coefficient of the adsorbed species is the average over all individual
ones induced by the surface field.
364 H. Frster
2.5.3
Change of Bandshape
Information on interaction can also be derived from the vibrational structure of
electronic bands which makes possible the evaluation of the symmetry change
of the electron cloud. In those rare cases where this structure is preserved, it can
be either displaced as a whole without alteration of the distances between the
vibrational subbands or shows changes like the appearance of vibrational modes
forbidden for the free molecule, induced by the unilateral surface field or
suppressed in cases with motional components perpendicular to the surface.
In addition, broadening and coalescence of the lines of a structured band due
to a blurring of the rovibrational levels by concomitant enhancement of the
averaged absorption can be observed.
2.5.4
Appearance of New Bands
The appearance of new bands, not observed with the adsorbent or with the
gaseous adsorbate, indicates severe interference with the electronic structure of
the admolecule by electron exchange with the active site or by bond scission
which leads to the loss of its identity.
2.6
Reasons for the Rare Application of Electronic Spectroscopy
Compared with IR spectroscopy the development of UV-VIS surface studies was

slower mainly (i) because of the complexity and inadequate development of the
theory of electronic spectra and (ii) due to methodological difficulties.
UV-VIS spectra are complex as they contain the fine structure from rovibra-
tional transitions. Upon adsorption this structure can only rarely be observed.
Due to the blurring of the rovibrational levels mainly broad unresolved bands
remain with the complete loss of structural information. The assignment of
bands to certain electronic transitions is carried out in correspondence with the
direction of the frequency shift or with the aid of semiempirical concepts. At
present there is no strict theory relating the spectral changes to the nature of
molecular interaction. New surface complexes and intermediate surface com-
pounds have to be determined by comparison with the spectra of the same
compounds in solution.
UV spectra are distinguished by their large extinction coefficients compared
to the infrared. The methodological limitations are on one hand due to the fact
that the UV bands of most organic compounds lie in the far ultraviolet and this
region is not accessible using commercial spectrometers. On the other hand,
strong scattering in the ultraviolet region by powders is a serious drawback, so
that transmission experiments are frequently replaced by reflection studies. One
way to improve the information is combination with other methods like IR spec-
troscopy in order to obtain information on the changes in the vibrational struc-
ture, or electron spin resonance (ESR), which is especially valuable in the case of
radicals.
3
Experimental Techniques
3.1
Absorption/Reflection Spectroscopy
Spectroscopy of molecularly dispersed systems is routine both instrumentally

and theoretically while that of heterogeneous systems like zeolites is still
nonroutine. Thus, it requires much more effort and this will be concisely outlined
in this section.
3.1.1
Dispersive vs. Nondispersive Spectrometry
Since the triumphant advance of interferometers in the infrared range, spec-

trometers are nowadays distinguished into interferometers or non-
dispersive spectrometers and dispersive instruments. There are two main
arguments for proving the superiority of interferometry over dispersive spec-
troscopy:
1. the throughput or Jacquinot advantage, i.e., the larger light gathering power
due to the lack of slits resulting in larger signals, and
2. the multiplex or Fellgett advantage, i.e., the simultaneous recording of the
optical signals giving rise to an improvement in the signal-to-noise ratio by
the square root of the number of resolution elements.
However, the latter is cancelled out when the noise is dependent on the square
root of the signal intensity (shot noise) as is the case with photomultiplier tubes
(PMTs) or can be reversed into a multiplex disadvantage for those cases where
the noise is directly proportional to the signal intensity (flicker noise).As the less
structured broad bands in the UV-VIS, especially with adsorbed species, do not
require the high resolution and the high wavelength accuracy (Connes advan-
tage) achieved by interferometers, an inferiority of interferometry over disper-
sive spectroscopy has been theoretically predicted, which, however, has to be
more carefully analyzed; this subject has been reviewed in more detail by
Williams [20].As a result, in the UV-VIS region mainly dispersive spectrometers
are used while interferometers are only rarely applied to the solution of special
nonroutine problems.
The basic components of dispersive spectrometers are a light source, a mono-
chromator, a detector with subsequent amplifying circuit and a readout/record-
ing device. The assembly of a traditional UV-VIS double-beam and, therefore,
ratio-recording spectrometer is shown schematically in Fig. 15 [21, 22]. Different
interchangeable light sources such as deuterium or tungsten-halogen lamps must
be alternatively used for the ultraviolet and the visible range, respectively, while
only one detector, the extremely sensitive PMT, is necessary. The light-dispersing
part can be designed as a single or a double monochromator depending on
whether one or two dispersing elements, today in general gratings, combined
366 H. Frster
Fig. 15. Schematic representation of the optical layout of a double-beam ratio-recording

spectrometer. Reprinted from [21]. Copyright John Wiley and Sons Ltd. Reproduced with
permission
with two or three slits are mounted together. Double monochromators furnished
with holographically ruled gratings ensure high resolution, light-gathering
power, wavelength accuracy and low stray light.
An interesting alternative is the photodiode spectrometer shown schemati-
cally in Fig. 16. The monochromator, sometimes also called a polychromator, has
only one entrance slit and no moving parts. The beam that is spacially resolved
by a fixed grating falls upon a linear photodiode array thus making possible mul-
tichannel recording and fast electronic processing.
The advent of computers of reasonable price and their application to spec-
trometers, spurred by the rapid development of the Fourier transform technique,
brought about a number of improvements that nowadays are integral parts of
modern spectrometers, such as microprocessor control, self-checking routines,
rapid scanning and visual display systems as well as opportunities for data
manipulation and reduction, thus augmenting sensitivity, precision and fastness
of data acquisition.
Data processing has evolved into an important procedure in modern spec-
troscopy. Baseline correction by polynomial fitting is one of the basic procedures.
Sharpening of the bands can be achieved by deconvolution in the Fourier domain
and signal-to-noise enhancement by digital filtering, least-squares polynomial or
Fourier smoothing [7, 21]. Derivative spectroscopy, which has been reviewed in
more detail elsewhere [23], facilitates the detection of overlapping bands, weak
side bands, resolution of shoulders as well as their more precise localization [24]
and can be carried out by computation, wavelength modulation or dual wave-
length operation. Dual wavelength spectroscopy is a versatile technique [7, 25]
that has also been successfully applied to zeolitic systems. In spite of the progress
obtained by data processing it should be kept in mind that good spectra will be
attained by a balanced application of an optimized parameter setting on the spec-
trometer and of data manipulation, letting mathematics solve those problems
Fig. 16. Schematic representation of the optical layout of a diode array spectrometer. Reprinted
from [21]. Copyright John Wiley and Sons Ltd. Reproduced with permission
that the optics cannot [21]. For further information on spectroscopic instru-
mentation, the reader is referred to the literature [7, 22].
3.1.2
Comparison Absorption vs. Reflection
In ordinary electronic spectroscopy the transmission mode is predominantly

applied, i.e., the absorbed light is measured as the absorbance given in Eq. (3)
related to concentration and sample thickness provided the Bouguer-Lambert-
Beer law holds over the entire range.
On powdered samples, however, the effects of reflection, refraction and dif-
fraction, summarized into scattering, cause severe losses in transmittance.
According to theory [26], the scattered intensity Is from an incident beam with
I0 depends on the wavelength l, the volume V and the number of particles per vol-
ume unit N and the relative refractive index n, derived for spherical particles is
Is = I0 24 p3 (n2 1/n2 + 2)2 NV2/l4 (21)
Equation (21) is valid for Mie scattering, i.e., if the particle sizes are larger than
the wavelength of light, as well as for the limiting case of Rayleigh scattering
where the particle size is smaller compared to l. On account of these accompa-
nying processes, transmission experiments with powdered samples are termed
as being carried out in scattered transmission.
Since according to Eq. (21) the scattered light increases with the fourth power
of the wavenumber, in the UV-VIS range it should be more advantageous to
replace the transmission by the reflectance method. However, Sendoda et al. [27]
compared both methods using nickel-exchanged X zeolite either packed as a
layer of 3 mm thickness for the reflectance or as pressed self-supporting wafers
368 H. Frster
for the transmission experiments. Their general experience was a good

agreement of the results obtained with both procedures, although the trans-
mission method yielded more intense bands. Encouraged by these findings
several groups decided to carry out their investigations in the transmission
mode [28, 29]. The advantages of the transmission technique are (i) a demand
for a lower sample amount, (ii) an easier sample handling concerning degassing
and admission of adsorbate, (iii) no need for an extra diffuse reflectance
accessory, and (iv) the omission of data conversion by means of the Schuster-
Kubelka-Munk function. A typical experimental setup for transmission studies
on zeolites is shown in Fig. 17 [30]. The zeolite powders are applied as self-
supporting wafers with a thickness between 510 mg/cm2 placed in a gold
foil frame and inserted into a quartz cell fused to a Conflat flange and connected
to a vacuum/dosage system. The heat treatment is achieved by surrounding
the lower part of the cell by a tubular oven while the spectra are recorded at room
temperature.
The vast majority of zeolite systems has been investigated applying reflectance
spectroscopy. Diffuse reflection is caused by single, multiple and dependent scat-
tering, the latter being a multiple scattering with phase coherence between the
scattered photons. The reflectance R is defined as the ratio of the total intensity
J0 of light reflected from a sample to the total intensity I0 incident
R = J0/I0 (22)
Fig. 17. Schematic representation of the experimental equipment for transmission UV-VIS
studies on zeolites. C Quartz cuvette; IM ionization gauge; CM capacitance manometer; IP ion
pump; TM thermal conductivity gauge; SP sorption pump; MP mechanical pump; 14 gas inlet
valves. From [30] reproduced by permission of The Royal Society of Chemistry
In the case of both scattering and absorption the amount of light backscattered
in a unit volume of the illuminated material is represented by the scattering co-
efficient S(l) and the part of light absorbed in the unit volume of the specimen
by the absorption coefficient K(l). Both effects have been unified in the Schus-
ter-Kubelka-Munk (SKM) equation:
F(R) = (1R)2/2 R = K/S (23)
where R is the reflectance of an infinitely thick sample. The SKM function

replaces the Bouguer-Lambert-Beer law and enables the determination of the
loss of absorbed radiation for strong scatterers. The theoretical background for
the derivation of Eq. (23) can be found in the literature [7, 26, 3134].
The schematic setup of a diffuse reflectance spectrometer which is obtained
by supplying a normal UV-VIS spectrometer with a diffuse reflectance accessory
is shown in Fig. 18. The core is an integrating sphere coated with a highly reflec-
tive material that collects the scattered radiation and leads it to the detector. The
light remitted by the sample is compared with that of a white standard as the ref-
erence. Zeolite layers of ca. 5 mm depth are generally sufficient to meet the con-
dition of infinite thickness. A typical sample cell furnished with an evacuated
Infrasil double window and designed for temperature control and evacuation is
depicted in the lower part of Fig. 18 [32]. Other cell constructions have been
reported by Klier [31] and Schoonheydt [34].
3.2
Photoacoustic Spectroscopy
Photoacoustic spectroscopy (PAS) is based on the photoacoustic or optoacoustic

effect: Provided modulated radiation impinges a solid in contact with a filler gas
in a closed volume, the absorbed light is transferred via heat into mechanical
energy as a sound wave registered by a highly sensitive microphone. This process
consists of a sequence of steps as illustrated in Fig. 2: Light absorption promotes
a system within 10151014 s into excited singlet states S1, S2, Radiationless
deactivation from higher singlet states to S1 occurs within 10131012 s and from
S1 to the ground state within 109108 s under conversion into heat (IC).
Undesired for PAS are some competing processes like (i) radiative deactivation
within the singlet term system, (ii) intersystem crossing followed by phospho-
rescence, and (iii) photochemical reactions.
As shown in Fig. 19 for solid samples, monochromatic light, chopped at a fre-
quency in the order of magnitude of 101000 cps which is low compared with the
velocity of deactivation, strikes the solid sample contained in a sample holder.
After excitation and relaxation the released heat diffuses to the surface, passes
into the gas phase and acts as an acoustic piston which generates a pressure wave
detected by the microphone and amplified by a phase-sensitive amplifier locked
to the chopping frequency w. Solution of the heat diffusion equation proves that
after a distance x from their starting point the heat waves are damped by:
exp (x/s) (24)

370 H. Frster
Fig. 18. A Schematic representation of a diffuse reflectance spectrometer. For (i) nondiffuse dis-
persed illumination the source at the left and the detector at the integrating sphere are used.
For (ii) diffuse nondisperse illumination (in the case of fluorescent samples) a source directly
attached to the sphere and the detectors on the left upper side are applied. B Temperature-reg-
ulated and evacuable diffuse reflectance sample cell according to [32]. Reprinted from [32] with
permission of Academic Press, Inc
Fig. 19. Schematic representation of a photoacoustic cell for solid samples. Reprinted from [7]
with ms being the thermal diffusion length, i.e., the length within which the heat
wave is diminished by 1/e. ms is related to the thermal diffusivity a and the
chopping frequency by:
s = 9
2a/w (25)
The PA signal is directly proportional to the absorption coefficient:
b = (1/d) ln I0/I (26)
where d is the pathlength of the sample and thus is proportional to the number
and energy of the photons. Therefore, this method is particularly suited for mea-
surements in the UV-VIS region. The PA signal is usually normalized to that of
a black body approximated by a carbon standard as the reference. For more
details of the origin and interpretation of the complex PA signal the reader may
consult the literature [7, 32] and the references given therein.
The advantages of this method against other techniques are:
1. PAS is well suited for the measurement of opaque materials.
2. As according to Eq. (25) the thermal diffusion length ms is inversely propor-
tional to the square root of the chopping frequency w, nondestructive depth
profiling is possible.
3. There are minimal demands to sample preparation.
4. The detector is insensitive to light and therefore no interference with the emis-
sion of radiation by the sample will take place.
Comparison of the PAS technique with the diffuse reflectance spectroscopy
(DRS) technique has been discussed [7, 32].
372 H. Frster
3.3
Luminescence Spectroscopy
In contrast to PAS this method makes use of the radiative deactivation upon exci-
tation by fluorescence or phosphorescence, as illustrated in Fig. 3. According to
the Kasha rule fluorescence starts from the first excited singlet state S1, while
phosphorescence is the result of ISC followed by a spin-forbidden T1S0 tran-
sition, both reflecting the vibrational structure of the electronic ground state.
As nonradiative deactivation also has to be taken into account for the fluo-
rescence intensity F, the general expression of which is more complicated and can
be found elsewhere, e.g., [7], the Bouguer-Lambert-Beer law does not always
hold. In the simplified case of low concentrations c and extinction coefficients e,
F is directly proportional to c and e. While the latter relates F to the absorption
spectrum, the former gives rise to quantitative analytical application.
Luminescence spectroscopy is a tracer technique mostly applied to first-row
transition metal and rare-earth ions. Its great advantage is the very high sensi-
tivity that, for zeolites, facilitates the study of isomorphously substituted frame-
work as well as exchangeable cations. However, the high sensitivity can also be a
disadvantage, as spurious signals due to lattice defects may be observed. In addi-
tion, too high a concentration may give rise to concentration quenching, and the
emission is dependent on the symmetry of the coordination sphere, so that for
many applications luminescence is not quantitative.
The schematic layout of a luminescence spectrometer for recording of fluo-
rescence, phosphorescence and excitation spectra is shown in Fig. 20. As excita-
tion source a mercury vapor discharge lamp or, as a continuous source, a pulsed
xenon flashlamp is used. Excitation spectra are obtained by setting monochro-
mator b2 at the maximum of the luminescence spectrum and varying by means
of the monochromator b1 the exciting wavenumber continuously over the whole
absorption spectrum of the sample. The luminescence spectrum is corrected by
diverting a portion of the incident light on a beamsplitter d to a quantum counter
Fig. 20. Scheme of a fluorescence spectrometer for excitation and luminescence spectra. a Light
source; b monochromator; c wavelength advance; d beam splitter; e quantum counter; f PMT;
g high voltage supply; h amplifier; i cuvette compartment; k ratiometer; l recorder. Reprinted
from [7] with permission of Springer, Berlin Heidelberg New York. Copyright 1992
e and dividing the amplified signals in the ratiometer k. The resultant lumines-
cence excitation spectrum represents the corrected intensity recorded as a func-
tion of the excitation wavenumber. For a more detailed discussion the reader is
referred to several books [7, 35, 36] that give an excellent overview of the state-
of-the-art luminescence techniques, the latter two with special reference to solids.
Applying lasers as the excitation source and either a scanning monochroma-
tor connected to a boxcar integrator or, better, an optical multichannel analyzer
for the experimental setup has given rise to the development of the laser-induced
fluorescence technique which can be used for diagnostic purposes in many con-
texts [37]. Pulsed UV lasers like nitrogen, frequency-tripled Nd:YAG or excimer
lasers serve as the light source.
Another luminescence-based method is luminescence lifetime measurement
in which the decay of the excited states with time is followed [38]. Luminescence
lifetime is the time required for the intensity to decrease to 1/e of its initial value.
The most common technique is time-correlated single photon counting which
uses a pulsed light source, one monochromator for the excitation and one for the
emission side, a PMT with fast response as the detector, a time-to-amplitude con-
verter and data storage in a multichannel analyzer.
Although earlier luminescence as a tool for probing catalysts and zeolites
found only modest application, there is growing interest in this method, as has
been documented in two reviews [39, 40].
4
Applications
Electronic spectroscopy has significantly contributed to the characterization of
zeolites and of molecules adsorbed on their internal surface. A comprehensive
review of UV-VIS spectroscopic investigations on zeolitic systems is too wide a
scope to cover in this article. Instead different topics of application will be sur-
veyed illustrated by selected examples taking into special account more recent
work. As in zeolite science it has become fashionable to speak of host/guest
chemistry, so first the host and later the guest will be discussed.
4.1
Characterization of the Host
4.1.1
Optical Properties of Zeolites
As SiO2, one of the parent materials of classical zeolites, is almost totally trans-
parent over the whole range between 55,50012,500 cm1, porous silica should be
expected to have similar optical properties. However, upon inspection of their
diffuse reflectance spectra, the differently structured zeolites MFI, LTA, FAU and
the mesoporous material MCM-41 revealed an onset of absorption near
35,000 cm1 increasing in intensity to higher frequencies [41].While the structure
type and the Si/Al ratio had no effect, impurities like iron or residual template
were decisive factors, the former due to a CT transition [42], the latter due to car-
374 H. Frster
bonaceous residues which, in insufficiently calcined samples, absorb up to 80%

of the incident UV light. Furthermore, the crystal size is of major influence as,
within the framework of the Schuster-Kubelka-Munk theory, the function F(R)
in Eq. (23) increases with growing particle size [26].
This absorption phenomenon observed for zeolites cannot be explained by
band gap absorption. As thermal treatment results in a reduced reflectivity and
a shift of the absorption onset to lower wavenumbers, it seems rather probable
to assign it to structural defects. Such intrinsic defects have been observed in
amorphous silicon dioxide by ESR as well as in crystalline silicalite by pho-
tothermal deflection spectroscopy [43]. Although the details are not yet fully
understood, there are at present the following speculations: An oxygen atom may
be set free from an -Si-O-Si- bond and forms an interstitial Frenkel defect. The
remaining -Si-Si- bond is split and a hole trapped by one of the silicon atoms
forming an E center, which is a silicon dangling bond. This idea is supported by
a paramagnetic line at g=2 in the ESR spectrum. The E center may react with an
oxygen molecule under formation of a dangling peroxide radical -SiO2 and/or
a peroxide bridge -Si-O-O-Si- in the framework.
From this hypothesis the recipe for preparing zeolites highly transparent in
the ultraviolet region should comprise both template-free synthesis and avoid-
ance of calcination. In this way a ZSM-5 sample has been synthesized lacking
absorption down to 50,000 cm1. Zeolites with such optical properties are well-
suited candidates for the development of luminophores. Luminescing guests like
rare earth or other ions are excited by the lines of the mercury plasma in the
ultraviolet region and emit radiation in the visible range; this phenomenon has
been proved by Tb3+ and Ce3+ exchange into zeolite A [41]. Broadening and split-
ting of the emission bands indicate the occupation of different cation sites. In
general, in addition to other properties, the exchange of the nonframework
cations influences essentially the luminescence behavior of zeolites depending on
the zeolite type, the reducing agent and the water content of the sample [44, 45].
In the last two decades there has been considerable progress in the synthesis
of molecular sieve materials. One result of these efforts is cloverite, a large-pore
gallophosphate molecular sieve with a 20 T-atom cloverleaf-shaped entrance into
a supercage of about 30 diameter, which has attracted attention not only due
to potential novel adsorption and catalytic properties and the uptake of bulky
adsorbates, but also as a host for the preparation of semiconductor nanoclusters,
as an example for the development of new advanced materials.
For this intention cloverite had to be thoroughly characterized by a host of
analytical tools, among others by UV-VIS DRS in order to assess its optical prop-
erties and their change upon dehydration and calcination in oxygen at 400 C
[46]. As shown in Fig. 21, the spectra of the as-synthesized and slightly dehy-
drated samples reveal an OGa/P band-gap excitation near 250 nm. After heat
treatment at 350 C, four bands between 230 and 340 nm arise, somewhat blue-
shifted upon treatment at 400 C, which then merge into a broad intense absorp-
tion centered around 300 nm. For assigning these bands there are at present only
speculations, the most probable of which attributes them to pp* transitions of
occluded polycyclic aromatics originating from the pyrolysis of quinuclidine
applied as template. A certain similarity to C60 buckminsterfullerene cannot be
Fig. 21. UV-VIS spectra of cloverite samples. A As-synthesized; B dehydrated at 170C;

C 350C; D 400C; and E calcined at 400C. Reprinted with permission from [46]. Copyright
1993 American Chemical Society
ignored. In this temperature range, the simultaneous depletion of cloverite from

frame-work hydroxyls by condensation and residual template by cracking can be
established.
4.1.2
Siting, Oxidation State, Coordination Sphere of Framework and Nonframework Cations
The properties of zeolites are decisively dependent on their cations, either on the
framework cations introduced by replacement of Si or Al in the zeolite lattice or
on the charge-compensating nonframework cations inserted by ion exchange in
the solid state or from solutions. This opens possibilities for targeted modifica-
tion of these materials in order to tailor their properties for specific applications.
The predominant importance of the cations in zeolites is that they form so-
called active sites for selective interaction with guest molecules in sorption and
catalytic processes. From the point of view of advanced material science [47] they
play a significant role in the formation of quantum-sized clusters with novel opti-
cal or semiconducting properties. As they give rise to cationic conductivity, zeo-
lites can be used as solid electrolytes, membranes in ion-selective electrodes and
as host structures in solid-state batteries. Organometallic compounds and coor-
dination complexes can be readily formed on these cations within the larger
cages or channels and applied to gas separation, electron-transport relays and
hybrid as well as shape-selective catalysis [48].
In the case of framework cations the T ions are isomorphously replaced by
other ions which may be directly accessible due to the internal micropore surface.
376 H. Frster
However, they can also, if they are hidden, influence or change the overall prop-
erties of the zeolite framework.
The exchangeable nonframework cations are distributed over all potential
sites such that the best charge balance is achieved and the energy of the unit cell
becomes minimal. For detailed information the reader is referred to an atlas com-
piling among other things the extraframework sites of 36 structure types [49].
One principal difference to framework cations is that the former are mobile,
which on the other hand means that one can never work on the assumption that
the cations stick in fixed positions. On the contrary, they undergo a dynamic
motion as in solution, jumping from site to site within the lattice. In addition an
induced motion in the presence of ligands, other types of cations or hydroxyl
groups, i.e., on all occasions that modify the charge distribution at the sites, has
been established.
The crystallographic sites of nonframework cations in the structures of zeo-
lite A (LTA) and zeolites X and Y (FAU) are indicated in Fig. 22 where their coor-
dination to the framework is shown in the lower part. Their nomenclature, sites
per unit cell, site symmetry and location are summarized in Table 3. Note that the
SI position in LTA is equivalent to site SII in FAU and will be further designated
as SIA.
Fig. 22. Schematic representation of the structures of zeolite A (LTA) left and zeolites X and Y
(FAU) right with indication of the cation sites. In the lower part the coordination of the cations
to the framework is displayed. Reprinted from [58], p 133, by courtesy of Marcel Dekker, Inc
Table 3. Nomenclature, number per unit cell (u.c.), symmetry and location of nonframework
cation sites in zeolites LTA and FAU
Zeolite Site Number of sites/u.c. Site symmetry Location
LTA I 8 C3v 6R, a cage

II 3 C1 8R, a cage
III 12 C2v? 4R, a cage
FAU I 16 Oh D6R, g cage
I 32 C3v D6R, b cage
II 32 C3v 6R, a cage
II 32 C3v 6R, b cage
III 64 C4v 4R, a cage
The considerable efforts that have been expended to discover location, coor-
dination, ligancy, oxidation state, stability, and migration of the cations and the
details of their interaction with guest molecules are documented by the vast
number of publications. Different experimental methods have been used for this
purpose, e.g., X-ray diffraction (XRD), extended X-ray absorption fine structure
( EXAFS), IR, UV-VIS and ESR spectroscopy, calorimetry, thermogravimetry, dif-
ferential thermal analysis, dielectric and magnetic measurements, to mention
just a few, as well as their combined application. Nowadays, increasing support
is given to them by theoretical methods such as quantum mechanical calcula-
tions, lattice energy and free energy minimization together with computer graph-
ics [50].Although from all these techniques XRD should be the method of choice,
there are still some drawbacks, as sufficiently large single crystals are frequently
not available and powder patterns are rather complicated so that Rietveld refine-
ment is required. However, this does not alter the fact that, at low cation concen-
trations, this method becomes unreliable per se. In this case UV-VIS spec-
troscopy together with other techniques has contributed greatly to the
enlightenment of this topic and has to be considered as its most important appli-
cation. The results have been summarized in a number of review articles [3134,
5158].
4.1.2.1
General Aspects of the Influence of the Environment on the Electronic
and Optical Properties of Transition Metal Ions in Zeolites
For investigations in the visible range, transition-metal ions (TMI) are excellent
probes for the study of cation siting, coordination, ligancy, oxidation state and
migration in zeolites. Knowledge of the introductory literature dealing with the
theory of bonding in TMI complexes and with the theoretical interpretation of
their spectra, e.g., [59, 60], seems to be appropriate for a better understanding.
With respect to the TMI the zeolite framework behaves like an anion, a solvent
or a ligand. The ions are usually introduced from aqueous solution as hexaquo
floating complexes [M(OH2)6]n+. Upon partial dehydration they become stabi-
lized as wall complexes bound in part to the framework and in part to remain-
378 H. Frster
ing water. Upon complete dehydration they become complexed as bare ions only
to the framework. In zeolites A, X and Y the dominant site for trapping the cation
is the distorted hexagonal oxygen window (6R) that is of D3h site symmetry. Klier
et al. [52, 56] have developed a theoretical model for this site.As shown in Fig. 23,
the three basal ligands, L1L3, are formed by the three proximal oxygens of the
6R, while the fourth ligand L4 represents either a sorbed guest molecule or the
axial electric field from the zeolite framework charge. The ligand field for this
model is of C3v symmetry which reduces for L1=L2=L3, L4=0 and b=90 to D3h and
for L1=L2=L3=L4 and b=109.48 to Td. The Hamiltonian for a generalized ligand
field at a site of the symmetry group G has the form [52]:
H = S [(h2/8p2 m) i2 Z e2/ri +VG (ri , qi , ji)] + S e2/rij + HJT + HLS (27)
where in the square brackets the first two are the kinetic energy and the nuclear-
electron terms and VG is the ligand field operator (G=C3v, D3h, Td) representing
the potential experienced by an electron on the central ion as a result of the
repulsion by the ligands and HJT and HLS are the Jahn-Teller and spin-orbit cou-
pling contributions. Taking at first only interelectronic repulsion Se2/rij and lig-
and field interaction into consideration, from diagonalization of their respective
matrices the electronic energies and wavefunctions are obtained.While the elec-
tron repulsion matrix elements are evaluated in terms of the Racah parameters
B and C, the matrix elements of VG are evaluated in terms of two parameters, G2
and G4, which account for distribution and strength of the ligand field G. G2 and
G4 depend on each other, and for a given charge distribution G2/G4 is kept con-
stant. Using G2 and G4, the electronic wavefunctions are parameterized by select-
ing values that best fit the calculated to the spectroscopically observed electronic
energies. For a D3h complex, e.g., the splitting of the orbitals becomes:
D1 = E(e) E (a1) = dxz VD3h dxz dz2 VD3h dz2

= (1/28)[6G2 15 G4] (28)
and
D2 = E(e) E(a1) = dx2y2 VD3h dx2y2dz2 VD3h dz2

= (1/28) [24G2 15G4/2] (29)
which for an assumed value of G2/G4=10 reduces to D1=(45/28)G4 and

D2=(465/56) G4. These splittings are large compared to those familiar splittings
in tetrahedral or octahedral fields Dt=(60/81)G4 and Do=(5/3)G4.
The right hand side of Fig. 23 it is shown how the one-electron energies of a
free TMI change when it is first brought into a spherical field and then fields of
D3h, C3v and Td symmetry are switched on. The transition from D3h, through C3v
to Td is accompanied by an increase of L4 from 0 to 1 and of b from 90 to 109.48,
shown in the lower part of Fig. 23.
The energy term diagrams, i.e., energies of multi-electron states vs. parame-
ter G4 both in units of Racah parameter B, obtained for dn+5(dn) ions in D3h sites
UV/VIS Spectroscopy
Fig. 23. C3v ligand field model for intrazeolite transition-metal ion complexes (left). Changes of the one-electron energy levels starting from the free ion
and switching on fields of spherical, D3h, C3v and Td symmetry (right). Reprinted from [32] with permission of Academic Press, Inc
379
380 H. Frster
(b=90, G2/G4=10, C/B=4) [56] are shown in Fig. 24. These diagrams can be used
for the assignment of the main transitions in UV/VIS spectra, as will be exem-
plified in Sect. 4.1.2.3. However, for a detailed interpretation of the fine structure,
a refinement of the ion position and of the term splitting due to the Jahn-Teller
effect and spin-orbit coupling is necessary.
The D3h position of a TMI is a very stable one. Provided the ion moves along
the trigonal axis under transition to C3v symmetry, there will be no removal of
degeneracies and only a small energy change upon increasing the angle b by 30.
This justifies the application of the D3h ligand field also to near-planar complexes.
Jahn-Teller distortion, on the other hand, arises from off-axial motions of the
cation [56].
When the trigonal site is approached by a guest molecule (L4 in Fig. 23) and
the strength of interaction with L4 is comparable to that with the proximal
framework oxygens L1L3, there are considerable changes in the energy levels
and triply degenerate states T1=A2+E and T2=A1+E of the Td group will be
formed which, in the case of symmetry lowering, will split into symmetrical
triplets.
Furthermore, from these diagrams, a number of useful properties can be
obtained, e.g.:
1. ligand field stabilization energies
Ec = Egs (G4)Egs (G4 = 0) (30)
with Egs the energy of the ground state,

2. one-electron oxidation potentials DE(1) for oxidation processes of the kind
M2+ (D3h) + O2 M3+ O2 (Td) (31)
3. two-electron oxidation potentials DE(2) for oxidations
M2+ (D3h) +1/2 O2 M4+ O2 (Td) (32)
4. reduction potentials DE(3) for reduction to neutral atoms
M2+ (D3h) + H2 Mo + 2 H+ (33)
and
5. reduction potentials DE(4) for reduction to bulk metal according to
DE(4) = DE(3) DUcoh (34)
with DUcoh the cohesive energy of the metal.

The quantities (1)(5) have been calculated for G2/G4=10 and G4/B=3; they are
depicted in Fig. 25 and are given here without further discussion.
Other theoretical approaches for the analysis of spectroscopic data and the
evaluation of the site preference are the angular overlap model (AOM) and the
Fig. 24. Energy term diagrams for dn+5(dn) ions in a D3h ligand field with the parameters b=90,
G2/G4=10, C/B=4. Reprinted with permission from [56]. Copyright 1977 American Chemical
Society
382
Fig. 25. Ligand field stabilization energies Ec, reducibilities to metal atoms DE(3) and bulk metal DE(4), one-electron DE(1) and two-electron oxidizabilities
DE(2) for divalent ions of the first transition-metal series in sites of D3h symmetry. Reprinted with permission from [56]. Copyright 1977 American Chemical
H. Frster
Society
electronegativity equalization method (EEM). The basic ideas of these concepts

may be inferred from [58]. For the details of calculation the original literature
cited therein should be consulted.
A combination of mathematical and chemometrical techniques has been pro-
posed by Verberckmoes et al. [61] for unravelling the contributions of differently
coordinated TMI to DR spectra such as Co2+ in zeolite A. Subsequently to the
decomposition of the spectra in Gaussian bands with the commercial software
package Grams 386, the chemometrical methods PCA (Principal Component
Analysis) and SIMPLISMA (SIMPLe-to-use Interactive Self-Modeling Analysis)
were applied to receive the spectra and intensities of the pure components. In this
way the DRS of CoA dehydrated at 400 C was resolved into two components of
trigonal and pseudo-tetrahedral symmetry.
4.1.2.2
Framework Cations
The first examples are devoted to zeolite synthesis (cf. also Vol. 1, Chap. 7 of this
series Molecular Sieves Science and Technology). Aluminophosphate-based
molecular sieves (acronym AlPO4) contain equimolar amounts of Al and P in lat-
tice T sites that can both be replaced by other elements. In the case of Co inser-
tion the acronym is changed to CoAPO or more generally to MeAPO. CoAPO is
usually synthesized using Et3N as template.
Isomorphously substituted tetrahedral cobalt can be proven by two triplets,
one (542, 580, 624 nm) in the visible and another (1329, 1492, 1719 nm) in the NIR
region and can be clearly distinguished from octahedral Co2+ (214, 304, 470 sh,
519 nm) [62]. Co2+ can be used as a sensitive probe for following its fate during
synthesis from the gel to the final product. At low template concentration most
of the Co2+ remains as octahedral [Co(H2O)6]2+ (519 nm) in solution with a small
amount of an octahedral complex [CoO4(H2O)2] (470 nm, sh) adsorbed on the
surface of pseudoboehmite gel particles and leading to CoAPO-5 (AFI topology).
At higher template concentrations tetrahedral surface complexes are formed
giving rise to the formation of CoAPO-n with chabazite topology.
The distribution of cobalt within the CoAPO-44 crystallites has been investi-
gated by photoacoustic spectroscopy making use of the depth-profiling capabil-
ity of PAS [63]. The triplet between 500650 nm in the spectra of the as-synthe-
sized sample (Fig. 26, left), being due to a transition from 4A2 to 4F(P), triply split
by Jahn-Teller distortion of tetrahedrally coordinated Co2+ in framework posi-
tions, becomes less resolved due to band broadening with decreasing modulation
frequency, i.e., increasing depth. In addition a band appears at 350 nm assigned
to a CT transition of cobalt complexes with cyclohexylamine used as the template
and obviously located in the inner part of the crystallite. The latter is also true for
nonframework cobalt species of possible lower symmetry and responsible for
spectral broadening at larger depths. On the contrary, after calcination the spec-
tra are independent of the frequency (Fig. 26, right), i.e., the cobalt distribution
has become more uniform.
Nonframework Co(II) in zeolites can only be partially oxidized. Substitution
in aluminophosphates changes this property to a usual one [64]. Upon calcina-
384 H. Frster
Fig. 26. Photoacoustic spectra of as-synthesized (left) and calcined CoAPO-44 at different
modulation frequencies: a 25, b 200 and c 1600 Hz. Reprinted from [63] with kind permission
of Elsevier Science NL, Sara Burgerhartstraat 25, 1055 KV Amsterdam, The Netherlands
tion of CoAPO in oxygen, the triplet mentioned above diminishes in favor of two
new bands at 320 and 400 nm; the latter attributed to LMCT transition in Co3+.
However, the remainder of the triplet indicates that part of the Co(II) cannot be
oxidized for which nonframework Co2+, non-stoichiometric CoO or resistant
framework species may be considered. Reduction with CO or H2 reverses the
spectral changes proving the reversibility of the redox reaction. Instead, water at
elevated temperatures can also be used as reducing agent.
Recently, NO has proven to be a valuable molecular probe for siting of frame-
work cobalt in CoAPO-18 [65]. From joint IR and UV-VIS investigations, the for-
mation of nitrosyl complexes at two different sites could be distinguished: (i)
Co2+ ions in the framework, associated with Brnsted acid sites as [Co2+(OH)P]
and (ii) structural defect Lewis acid Co2+ sites more slowly formed via activated
processes possibly requiring the cleavage of some framework Co-O bonds. The
change in the oxidation state of the cobalt ions was usually ascertained by DRS.
While, as mentioned above, the absorptions in the visible (500650 nm) and
near-infrared (10002500 nm) regions are typical of tetrahedrally coordinated
Co2+ ions, after admission of NO new bands appeared in the 220500 nm range,
dominated by an intense absorption at 260 nm, attributable to LMCT transitions
of NO complexes on Co3+ ions. The latter are presumably formed from
[Co2+(OH)P] by oxidation with NO at room temperature.
Titanium has been introduced into zeolite frameworks in order to prepare cat-
alysts which are both active and selective for oxidation of hydrocarbons with
H2O2. Catalysts of this type are suited for fine chemical production, as product
waste will be minimized. Starting materials for Ti-modified zeolites were ZSM-
5 (TS-1), ZSM-11 (TS-2), ZSM-48 and zeolite beta. Echchahed et al. [66] investi-
gated the synthesis of TS-1 under variation of hydrothermal time and tempera-
ture. The change in the UV band with maximum at 208 nm, typical of tetrahedral
Ti, which broadens with growing titanium incorporation, is explained by the
presence of two different framework Ti species: (i) open sites [Ti(SiO)3OH] and
(ii) closed sites [Ti(SiO)4]. An increasing number of the latter for higher load-
ing cause strong local stress and crystal fragmentation, while the former, due to
a bond rupture within the Si-O-Ti bridge, release the local tension.
Incorporating Ti into zeolite beta was carried out by hydrothermal synthesis
and by post-synthesis modification with ammonium titanyl oxalate [67]. Evi-
dence for isolated Ti(IV) most likely in a framework position was also given here
by the strong 212 nm band. Post-synthesis treatment with nitrogen saturated
with TiCl4 vapor was also successful in isomorphous incorporation of Ti,
enhanced by preceding acid leaching as revealed from the UV-VIS spectra [68].
Al-free Ti-beta preparation by dry gel conversion has been reported by Hari
Prasad Rao et al. [69]. The intense band in the 205225 nm region proves the
incorporation of Ti into the framework. A weak band around 270 nm and the
onset of absorption at 330 nm, however, suggest the presence of Ti at distorted
positions or as extraframework titanium, e.g., as oligomeric hexacoordinated Ti
species.
Corma et al. [70] attempted to insert Ti into the mesoporous structure MCM-
41 by direct synthesis. The UV-VIS spectra revealed strong absorptions in the
190300 nm range that confirm the absence of bulk anatase (330 nm) in agree-
ment with Franke et al. [71]. However, the presence of isolated Ti atoms in octa-
hedral coordination is proved by the LMCT transition near 225 nm which after
dehydration shifts to 205 nm characteristic for tetracoordinated Ti(IV) in TiO4
and TiO3OH environments. Additionally, a weak band around 270 nm points to
hexacoordinated Ti-O-Ti species which probably belong to an extra-wall amor-
phous phase. The synthesized Ti-MCM-41 turned out to be a catalyst of good
performance in oxidation of bulky olefins such as a-terpineol or norbornene.
Using the same Ti source, a Ti-MCM-41 synthesis via an acidic route has been
presented by Corma et al. [72] in which the samples obtained show a very nar-
row band of tetrahedrally coordinated Ti at 220 nm and missing higher-wave-
length shoulders, as found in samples prepared under alkaline conditions, point-
ing to a more distorted environment.
Unfortunately, Ti(IV) forms hydroxides of low solubility. In order to avoid pre-
cipitation during the synthesis of Ti-MFI the admixture of complexing agents like
acetylacetone has been studied [73]. Due to the DR spectra acetylacetone admis-
sion resulted in a higher amount of framework titanium (212 nm), again in the
absence of anatase, and in a distinct reduction of extraframework Ti species, e.g.,
like octahedrally coordinated titanium (274 nm).
Successful application of DRS together with complementary methods like
ESR, XRD, XAS, IR, SEM, NMR etc. for characterization of the synthesized micro-
porous materials with respect to incorporation into the framework has been
further reported for MFI-type vanadium silicalites [74, 75], for VAPO-5,
MgAPO-5 and MgVAPO-5 [76], for MFI-type ferrisilicalite [77] and for insertion
of chromium in a number of molecular sieves [78]. In the case of the latter, essen-
tially nonframework species could be identified and even in Cr-silicalite and
CrAPO-5 Cr3+ showed no tendency to change from octahedral to tetrahedral
coordination.
386 H. Frster
4.1.2.3
Nonframework Cations
As mentioned above, chromium in hydrothermally synthesized Cr-silicalite and

CrAPO-5 showed no motivation to change its coordination to tetrahedral which
is the requirement for isomorphous substitution into the zeolite framework. Evi-
dently, only a limited number of elements are able to form framework cations,
whereas, for nonframework cations, there are theoretically a vast number of
possibilities in various symmetries. Since nonframework cations are widely
accessible they can form complexes with different guest molecules. Their elec-
tronic spectra are fingerprints of the electronic structure of these entities and
give valuable hints to their identification. Distortions of the geometry and
vibronic coupling give rise to a fine structure, and, from their interpretation,
information on the stability of molecular species and, furthermore, on their reac-
tivity can be obtained. First, siting, local geometry, oxidation state and migration
of these ions in the absence of adsorbates will be considered. Water will be
excepted since, due to its permanent availability and due to the hygroscopicity of
zeolites, its presence must always be taken into account.
The classical investigations of TMI-exchanged zeolites, especially of the 3d ele-
ments, started on A-type zeolites exchanged with Ni2+ by Klier et al. [31] and were
later extended to the cobalt and chromium forms, see, e.g., [51, 52, 54, 57]. As the
optical spectroscopy of nonframework transition-metal ions has been reviewed
in several contributions [16, 31, 34, 51, 52, 54, 55, 57, 58], only an outline will be
given here and the main features will be considered briefly.
The optical spectra of fully hydrated zeolites largely match those of the hexa-
quo complexes, except the hypsochromic shifts of ca. 1500, 650 and 500 cm1 in
Cr(II)-, Co(II)- and Ni(II)-exchanged zeolites A [52]. These shifts might be
explained by a greater ligand field of the zeolitic water which, compared to nor-
mal water, is more structured, but the influence of the framework cannot be
excluded either. The similarity to the spectra in aqueous solutions indicates the
presence of ions octahedrally coordinated by zeolitic water. These floating com-
plexes are supposed to reside mainly in the large cages, although Seff reports
hexaquo cobalt(II) located in the center of the smaller sodalite cage and forming
twelve H-bonds to the framework of zeolite Co4Na4A [53].
After complete dehydration, the transitions from the Cr(II) (d4) ground
state 5E to 5E and 5A1 excited states (12,300, 17,000 cm1), from the Co(II) (d7)
ground state 4E to 4E, 4A1 (accidently degenerate with 4A2) and 4E excited
states (7000, 16,000, 25,000 cm1) and from the Ni(II) (d8) ground state 3A2 (F)
to 3A1 (accidently degenerate with 3A2),3E (~3E), 3A2 and 3E excited states
(4500, 9090, 22,760, 36,000 cm1) prove trigonal coordination (D3h) to the three
proximal oxygens Of of the 6R in site S1A of zeolite A as bare ion [M(Of)3]2+
which can be assigned using the energy term diagrams in Fig. 24. Although,
e.g., for nickel only the transitions to 3A1 and 3E are symmetry-allowed, the
appearance of the other two indicate a breakdown of the electric dipole transi-
tion selection rule even for small deviations from planarity [52]. The spectral
changes upon dehydration of Co(II)-exchanged zeolite A are shown in Fig. 27
[32]. The d-d band frequencies of some bare 3d TMI in zeolite LTA [52, 79]
Fig. 27. DR spectrum of fully hydrated zeolite Co(II)A (1) compared to that of an aqueous
Co(II) solution (2). Upon complete dehydration the spectrum drastically changes to (3)
assigned to a Co(II) ion in trigonal planar D3h coordination by the 6R. Reprinted from [32] with
permission of Academic Press, Inc
and their ground and excited electronic states in D3h ligand fields are compiled
in Table 4.
Dehydration is accompanied by gradual migration of the ions into sites in
which they become stabilized as wall complexes or bare ions. This process is
strongly dependent on the zeolite type, on the ion and on the dehydration con-
ditions. In zeolite CrNaY, e.g., it was shown by electronic spectroscopy that upon
dehydration with increasing temperature the Cr(II) ions become localized first
in SII sites, then in SII or SI sites and finally in Oh geometry in SI sites [57]. The
red shift of this last complex (14,300, 18,800 cm1) compared to the floating hexa-
quo complex (16,400. 23,600 cm1) might be explained by the weaker ligand
zeolitic oxygen compared to the oxygen of water.
Table 4. Sites, d-d band positions of some bare 3d TMIs in zeolite LTA and their ground and
excited electronic states in D3h ligand fields [52, 79]
3d TMI Site d-d Bands/cm1 Ground state Excited states
Cr2+ I 12,300, 17,000 5E 5E, 5A

1
Co2+ I 7000, 16,000, 25,000 4E 4E, 4A , 4A , 4E
1 2
Ni2+ I 4500, 9090, 22,760, 36,000 3A
2
3A , 3A , 3E, 3E,3A , 3E
1 2 2
Cu2+ I 10,400, 12,200, 15,000 2E 2E, 2A
1
388 H. Frster
Whereas the early investigations concentrated mainly on bare ion complexes

in LTA and FAU, later more and more different types of zeolites, like chabazite,
mordenite [80], MFI boralite [81] and nitrido zeolites [82], were investigated. The
latter, e.g., P-N-sodalite with the general formula M7-xH2x[P12N24]X2-y, where X
stands for halogen and M for a 3d TMI, attracted attention as they represent a
new class of zeolites with substitution of the anion substructure, in this case
replacement of oxygen by nitrogen. Features of these sodalite-like materials are
high thermal and chemical stability as well as intense coloration. Responsible for
the color is a near-tetrahedral coordination more likely approximated by the
complex [MX(NH3)3]+ which gives rise to an electronic spectrum similar to that
of [MX4]2, however, taking into account symmetry lowering and LS coupling.
A study of the Cu(II) siting in dehydrated CHA and MOR by means of DRS
assisted by ESR gave three d-d bands for CuNaCHA (10,700, 12,900, 14,800 cm1)
accompanied by a LMCT transition at 40,000 cm1 and for CuNaMOR (12,500,
13,700, 14,800 cm1) together with the corresponding LMCT transition around
42,000 cm1 [58, 80]. The spectra of CuNaCHA are ascribed in analogy to those
of CuNaFAU to Cu2+ ions essentially coordinated to the hexagonal rings, while
those sited in the distorted hexagonal prisms do not significantly contribute due
to the lower occupancy and the smaller extinction coefficient of pseudo-octa-
hedral compared to trigonal complexes. The symmetry of the latter seems to be
lower than C3v, probably effected by LS coupling and Jahn-Teller distortion.
According to the d-d transitions in CuNaMOR, a fourfold coordination either to
site I (eight-membered ring) or to site IV (at the wall of the large channel) has
to be assumed.
The positions of the d-d bands of bare Cu2+ ions in different zeolites and sites
are summarized in Table 5. An overview indicates that they are site-dependent,
e.g., the two types of six-rings at SI (LTA) and SII (FAU) can be discriminated
from those at SI (FAU, CHA). On the other hand, in a first approximation, they
are independent of the structure type, the Si/Al ratio as well as ordering and the
type of the co-cation [58]. The TMI feels only the first coordination sphere rep-
resented by the nearest oxygens.
After assignment of the bands the ligand field parameters such as splitting,
Racah and nephelauxetic parameters can be determined from the frequencies.As
an example, with fully hydrated Ni(II)X zeolite, five bands are observed at 5100,
6850, 8450, 14,700 and 24800 cm1 from which the first two have to be attributed
Table 5. Electronic band positions of Cu(II) coordinated to lattice oxygens in different zeolites
and sites, after [58], in brackets after [79]
Zeolite Site d-d Bands/cm1
LTA I 10,400 12,20012,800 15,00015,400

FAU II 10,400 (10,400) 12,40013,000 (12,400) 14,70015,100 (14,700)
FAU I 10,800 (10,700) 12,40013,000 (12,900) 14,70015,100 (14,800)
CHA I 10,700 12,900 14,800
MOR I 12,500 13,700 14,800
MOR VI ill-defined
to the n2+n3 combination and the 2n3 overtone band of adsorbed water [27]. The
latter three are due to the d-d transitions of hexaquo Ni2+ ions in Oh symmetry:
n1 [3A2g(F)3T2g(F)], n2 [3A2g(F)3T1g(F)] and n3 [3A2g(F)3T1g(P)], increasing
in frequency in this order. In a similar way to that shown in Eq. (10), the term
energies can be related to the splitting Dq and the Racah parameter B which for
the frequencies lead to the expressions:
D q = (1/10) n1 (35)
and
B = (1/15) (n2 + n3 3 n1) (36)
The nephelauxetic parameter b is obtained from comparison of B (from Eq. 18)

with B (free ion) which amounts to 1041 cm1 in the case of Ni(II). The usual val-
ues of b fall in the range of 0.70.9, while its reduction indicates an increase of
covalent character in the bonds. The results calculated from the frequencies men-
tioned above are Dq=845 cm1, B=943 cm1 and b=0.91. These values proved to
be somewhat dependent on the degree of Ni2+ ion exchange but changed dra-
matically upon dehydration. Dehydration at elevated temperature gave rise to a
decrease of b to 0.30 and to an increase of Dq to 1110 cm1. The unusual value for
the nephelauxetic parameter hints to a completely different environment in cor-
respondence with a migration into SI sites and to a Ni-O bond with lattice oxy-
gen being more covalent compared to water as the ligand.
However, it should be stressed that electronic spectroscopy, even when com-
bined with complementary tools like ESR, does not unambiguously assign the
TMI to a distinct site. Here the assistance of theoretical methods like AOM and
EEM or others is indispensable. On the other hand, it gives reliable information
on the accessibility of the ions to adsorbates, as will be shown below.
Exposure of TMI-exchanged zeolites to various guest molecules frequently
gives rise to color changes which sensitively indicate spectral changes. They pro-
vide evidence of a competition between complexation to lattice oxygens and
complexation to nonframework ligands. In this way, a large variety of sometimes
rather unusual complexes can be formed in the zeolite interior; see Lunsford [54,
55] for reviews of the early efforts in this field.
What happens with the spectrum may be suspected from Fig. 27 in which the
25,000 cm1 band was indicative for a trigonal near-planar coordination of Co(II)
to the proximal framework oxygens of the 6R in zeolite A. As shown in Fig. 28,
admission of, e.g., nitric oxide leads to the disappearance of this unique band
accompanied by the appearance of a strong band at 19,000 cm1 and the onset of
a CT transition at >31,000 cm1 [83]. The diminishing of this band assigned to the
4E4E transition has already been observed earlier upon exposure to oxygen,
CO, CO2, ethylene oxide, N2O, H2O, NH3, imidazole, cyclopropane and olefins like
ethene, propene and the butene isomers [52, 8487]. At higher degrees of cobalt
exchange with trans-butene this band did not completely disappear indicating
that not all Co(II) ions were coordinated by the hydrocarbons [87].
The explanation for the spectral changes is, in a first approach, the change
from near D3h to C3v or near Td symmetry effected by the adsorbed guest. Finer
details concerning orientation and rotational behavior of the guest can be
390 H. Frster
Fig. 28. Electronic spectra of fully dehydrated Co(II)A zeolite (3.9 Co2+/unit cell) and upon exposure to NO in increasing steps
of pressure. The inset relates NO pressure to the occupation per cavity. Reprinted from [83] with kind permission of Elsevier
Science NL, Sara Burgerhartstraat 25, 1055 KV Amsterdam, The Netherlands
inferred from the fine structure of the bands. The spectrum of a tetrahedral
Co(II) complex corresponds to transitions from the 4A2(F) ground state to the
triply degenerate 4T1(F) and 4T1(P) excited states. A dynamic Jahn-Teller effect
should lift the degeneracy resulting in a symmetrical splitting, as shown on the
left-hand side of the vertical dotted line in Fig. 29. But in the case of coordinated
unsaturated hydrocarbons an unsymmetrical splitting of the 7000 and 17,300
Fig. 29. Term diagram of a Co(II) ion (i) in a Td environment with degeneracy lifting of the 4T1
states by a dynamic Jahn-Teller effect indicated as a symmetric term splitting (left-hand side
of the vertical dotted line); (ii) in an environment of lower symmetry (e.g., C1) with removal
of the degeneracy (right-hand side of the vertical dotted line). The best fit for Co(II)-ethylene
is obtained for G4/B=6.6 (vertical dotted line). Reprinted from [32] with permission of Acad-
emic Press, Inc
cm1 bands is observed which must be accounted for by a static symmetry per-
turbation arising from the charge distribution of the ligand. The off-z-axis com-
ponents of the charge distribution give rise to a C2v perturbation that lowers the
complex symmetry to C1 or C1v resulting in an unsymmetrical splitting into 4A1
and 4A2 states on the right-hand side of the vertical dotted line in Fig. 29. Pre-
requisite is an approach of the naked Co(II) ion by the ligand in p geometry, i.e.,
the hydrocarbon molecular axis is transverse to the threefold z-axis of the trig-
onal complex. While on the basis of theoretical considerations ethylene behaves
like a freely rotating ligand, decreasing ability to split the 17,300 cm1 band fol-
lows steric hindrance increasing in the order ethene < propene < cis-butene <
trans-butene [52]. Further examples of coordination of transition-metal ions
with mono-, bi- and polydentate ligands have been reported recently [79]. While
the identification of sites is not always unambiguous, optical spectra provide reli-
able data on the accessibility of the ions.
The nonframework ions, especially the TMI, frequently form the so-called
active sites responsible for the catalytic properties of zeolites. In these cases, coor-
392 H. Frster
dination and polarization, in the case of protons protonation, of the guest is fol-
lowed by transformation to a different compound as has been proven, e.g., by the
isomerization of cyclopropane and n-butenes over A-zeolites [87, 88].
Nowadays, NO decomposition is of great importance from an environmental
point of view. Investigating the suitability of zeolite CoA as a catalyst for this reac-
tion, electronic spectroscopy turned out to be a valuable tool for the determina-
tion of surface intermediates and strength of adsorption. As shown in Fig. 28,
starting from the bare site, sequential increase of the NO pressure yielded spec-
tra assigned to complexes with multiple NO coordination which upon evacuation
could be converted into the near tetrahedral mononitrosyl Co(II) complex. NO
adsorption proved to be strong but reversible [83].
Complementary to reflectance and transmission spectroscopy, luminescence
experiments are very useful for the study of TMI siting, resonance energy trans-
fer and photocatalytic decomposition, especially in the case of Cu-containing
zeolites. Basic investigations on the redox chemistry of Cu(II) ions in zeolite Y by
Strome and Klier proved after dehydration a preferred location of Cu2+ ions in
SI sites while those in SI sites, being a center of inversion, only contributed
slightly to the DR spectrum [89]. Upon hydrogen admission, Cu(II) was readily
reduced to Cu(I).As inferred from the DR spectra shown in Fig. 30, the Gaussian-
Fig. 30. DR spectra of hydrated Cu(II)Y (a), dehydrated Cu(II)Y (b) and the product of admit-
ting oxygen to Cu(I)Y, forming a Cu(II)(Of)3O2 complex on SII sites (c). The 18400 cm1 pho-
toemission of Cu(I)Y is shown as spectrum d. Reprinted with permission from [89]. Copyright
1980 American Chemical Society
shaped emission spectrum of Cu(I)Y centered at 18,400 cm1 (spectrum d) has

been assigned to the spin-forbidden 3E(3D3 3d94s)1A(1S0 3d10) transition in a
C3v ligand field. When Cu(I)Y was exposed to oxygen in order to be reoxidized,
the reflectance spectrum was different to that of Cu(II)Y (spectrum c in Fig. 30)
and the emission was quenched resulting in a decrease in intensity and reduction
of decay time in favor of a competing radiationless process. As shown in Fig. 30
(spectra c and d), the emission band overlaps with the absorption band of the
product of oxygen interaction identified as a near-tetrahedral Cu(II)(Of)3O2
complex in SII sites with an O2 ligand in p geometry. For the formation of this
complex, the Cu(I) cations had to move out of SI and SI positions to SII sites. This
overlap indicates an efficient resonance energy transfer between Cu(I) as emit-
ter center and Cu(II)(Of)3O2 as absorber, i.e., between sodalite unit and
supercage.
Cu(II)-containing high-silica ZSM-5 has attracted substantial attention as an
active catalyst for the removal of nitrogen oxides. Ebitani et al. proposed that
excitation of Cu(I) ions around 300 nm causes the photocatalytic decomposition
of N2O [90]. The excitation of this band gave rise to an emission centered at about
540 nm which was assigned to the radiative decay of an excited Cu(I) dimer,
(Cu+-Cu+)*, to its ground state. N2O admission efficiently quenched the lumi-
nescence, so that an electronic interaction between (Cu+-Cu+)* and the anti-
bonding MO of N2O had to be assumed which releases the decomposition of N2O
into N2 and O2. Furthermore, Cu+ luminescence after deconvolution has been
used for a semiquantitative determination of Cu coordination controlled by the
(-Al-O-(Si-O)n-Al-) sequence [91]. Of four possible sites, the most active in NO
decomposition proved to be the one adjacent to a single Al atom.
As they are relatively cheap and are produced in the form of almost ultravio-
let-transparent materials, rare-earth-doped zeolites have attracted growing inter-
est as substitutes for more expensive phosphors which, applied in fluorescent
lamps, should be able to efficiently convert UV into visible light. As shown by
Borgmann et al. [92], Eu3+-doped zeolite X excited with 254-nm radiation gave
only weak emission.Additional insertion of molybdate caused absorption of UV
radiation and subsequent transfer from the excited metalate LMCT state to the
Eu3+ 5D level increasing the quantum yield up to 7%. Upon thermal treatment at
1000 C a dense phase of the composition Eu4Si4Al8O24(MoO4)2 was formed,
mainly consisting of a rare earth sodalite with 25% quantum yield, as shown by
the emission spectra in Fig. 31; the quantum yield was improved to 38% upon iso-
morphous substitution of Eu3+ by Y3+. Finally, the direct synthesis of sodalites
from oxides gave a maximum quantum yield of 55% for the composition
(Eu0.4Y0.6)4Si4Al8O24(MoO4)2 due to the absence of any by-products.
4.1.3
Acid Strength of Zeolites
One of the outstanding properties of zeolites is their acidity. This is usually

divided into Brnsted and Lewis acidity that make them capable of being applied
in, e.g., cracking and shape-selective catalysis. Characterization of zeolite acid-
ity comprises the determination of its nature, amount and strength and is com-
394 H. Frster
Fig. 31. Left Part of the structure of Eu4Si4Al8O24(MoO4)2 sodalite. Right Emission spectra of
Eu3+-doped zeolite X treated with 10 wt.% (NH4)2MoO4 at 500 C (a) and of sodalite
Eu4Si4Al8O24(MoO4)2 obtained from the former by phase transformation at 1000 C (b). The
f-f transitions of Eu3+ are indicated. Reprinted from [92] with kind permission of the Materi-
als Research Society, Warrendale
monly accomplished using a number of techniques such as titration, adsorption,

IR spectroscopy of OH groups and basic probe molecules, TPD and NMR, and,
to a lesser extent, electronic spectroscopy. Nevertheless, aromatic hydrocarbons,
for example, can be readily ring-protonated under formation of carbocations that
are known to give characteristic absorptions in the UV-VIS region. As the ease
of protonation depends on both the acid strength of the host and the basic
strength of the guest, electronic spectra with probes of decreasing basicity should
allow determination of the acid strength.
As an example, the H-forms of ZSM-5, FAU, USY, MOR, dealuminated MOR
(DEAMOR) and rare earth ion-exchanged zeolite Y (REY) have been probed with
the vapors of in order of increasing basicity toluene, p-xylene, m-xylene, 1,2,4-
and 1,3,5-trimethylbenzene [93]. The formation of the corresponding cyclo-
hexadienyl cations of 1,2,4- and 1,3,5-trimethylbenzene on HMOR samples with
decreasing Al content (HM, HM D-1...HM D-3) is shown in Fig. 32. The differ-
ences in intensities can be explained by the different basicities of the probes from
which the former is protonated by strong and medium, the latter in addition also
by weak Brnsted sites, i.e., 1,3,5-trimethylbenzene grasps the total acidity. As a
result, dealumination of HMOR decreases the total number, but increases the
number of strong Brnsted sites. From semiquantitative evaluation the relative
strength of Brnsted acidity was determined and this increases in the order
HNaY<REY<USHY=HMOR<DEAHMOR, HZSM-5.
4.1.4
Polarity of Zeolites
The functions of zeolites are versatile: they may behave as an anion, as a ligand
or as a solvent. As a consequence of the latter zeolites should have solvent-like
Fig. 32. DR spectra of 1,2,4- (left) and 1,3,5-trimethylbenzene (right) on HMOR samples.
Reprinted from [93] with kind permission of Elsevier Science NL, Sara Burgerhartstraat 25,
1055 KV Amsterdam, The Netherlands
properties like, e.g., polarity, dipole moment, dielectric constant, etc. Since polar-
ity is a characteristic of a bulk phase, micropolarity should be the more suitable
term for this property when the extension of the phase is restricted to the
nanometer scale as in the case of zeolite voids. This property does not have pre-
cisely the same meaning as in solution but reflects the electric field within the zeo-
lite interior that is essentially provided by the cations. There have been some
attempts to probe this property by means of photophysical techniques applying
condensed benzenoid hydrocarbons, either making use of the intensity change of
vibronic bands (Ham effect), of the shift of the emission maximum or of the
changes in lifetimes in the fluorescence spectrum [94, 95]. As shown in Fig. 33 by
the emission spectrum of pyrene in zeolites KX and CsX, the intensity ratio of
higher transitions to that of the 00 transition of a vibronic progression might
serve as a measure of the environmental polarity. Compared to solvents the polar-
ities of the faujasite supercages are similar to those of methanol or acetonitrile.
As with solvents the intracavity formation of ionic CT complexes has been
exploited to give a description of the supercage polarity. The contact ion pair
methylviologen monoiodide undergoes a solvatochromic shift in acetoni-
trile/water mixtures due to the decrease in the energy gap between the ground
and the excited state with decreasing polarity, i.e., with decreasing water content
[96]. From comparison of the lmax position to that obtained at 362 nm in zeolite
Y the polarity of its supercage was estimated to be equivalent to that of a 50%
aqueous acetonitrile solvent.
Zeolites can also be considered as solid electrolytic solvents with a donor-
acceptor interaction between the framework and the counter cation. The color of
iodine is due to a p*s* transition in the visible range and is known to be
largely sensitive to the solvent [3]. Because of a donor-acceptor interaction
between iodine and the solvents the s* orbital is perturbed and shifted to higher
396 H. Frster
392 vs. 372 nm amounts to 0.49 and 0.60 as indicated on the spectra. Reprinted from [94] with permission of Else-
Fig. 33. Emission spectra of pyrene in KX and CsX zeolites. The intensity ratio of the monomer fluorescence at
vier Science, Inc
energies with increasing donor strength of the solvent resulting in a hyp-

sochromic shift of the corresponding band. Therefore, iodine should also be suit-
able for probing the donor strength, i.e., the basicity of zeolites. This idea has
been proposed by Choi et al., who systematically investigated the frequency shift
of the visible band in zeolites LTA, FAU, ZSM-5, LTL MAZ and MOR [97]. Upon
adsorption of iodine vapor at room temperature, a broad unstructured band was
observed, the maximum of which was shifted to shorter wavelengths compared
to iodine in CCl4 solution dependent on the zeolite type, the Si/Al ratio, the
counter cation and residual moisture. The probe does not adsorb on the cations
but forms a CT complex with the negatively charged framework oxygen atoms.
Moisture prevents iodine from a direct interaction and gives rise to a restricted
blue shift. For alkali-exchanged faujasites a linear negative correlation of the
absorption energy with the ionization potentials of the counter ions and the Si/Al
ratio, respectively, could be established. Due to the larger hypsochromic shift it
must be concluded that the FAU framework has a higher donor strength than LTA.
4.2
Characterization of the Guest
As already shown in the last section, electronic spectroscopy offers two possibil-
ities of characterizing a guest molecule encapsulated in the zeolite cage either (i)
an indirect one by following the changes of the host spectrum upon admission
of an adsorbate or (ii) a direct one by observing the spectrum of the guest. For
this case the adsorbate molecule must be capable of absorbing radiation in the
UV-VIS region accessible with commercial spectrometers. The latter will be in
the narrow sense the topic of the following section.
4.2.1
Inorganic Compounds
A number of simple inorganic compounds exhibit electronic spectra mainly aris-

ing from np*, pp*, p*s* and CT transitions in the UV range. From sul-
fur-containing compounds, sulfur dioxide as an example shows a less intense
band at 360 nm and a stronger band at 290 nm due to singlet-triplet and singlet-
singlet np* transitions [3]. SO2 is one of the source materials for the produc-
tion of elemental sulfur, accomplished by the modified Claus process in which it
reacts with H2S over alumina or bauxite [98, 99].Although at present zeolites will
not replace these materials in the industrial process, they can serve as useful
model catalysts [100]. Adsorbed on zeolite NaX, SO2 exhibits three bands at 215,
250 and 280 nm which, supported by IR investigations, are assigned to
chemisorbed HSO3, S2O5 and physisorbed SO2, respectively [101]. The second
reactant, H2S, adsorbed on the same adsorbent, gives rise to one band that shifts
with increasing coverage from 235 to 250 nm [102]. This electronic spectrum is,
again combined with IR experiments, explained at low H2S concentrations by a
complete dissociation of H2S into H+ and S2 on strong centers of interaction.
Increasing the coverage, formation of SH ions on weaker centers and finally
nondissociatively physisorbed H2S species were identified. For the elucidation of
the elementary steps of the Claus reaction both reactants must be brought
together. After preadsorption of H2S on NaX at room temperature the UV-VIS
band mentioned above appeared (Fig. 34, spectrum 1). Treating the sample in the
IR beam (spectrum 2) yielded SH species perceptible from the 2560 cm-1 band
shown as trace A in the inset of Fig. 34. After exposure to SO2 and subsequent
heating in the IR beam, the SH species disappeared (see inset, trace A) evi-
dencing their involvement in the reaction with SO2 together with the appearance
of new bands in the 250500 nm region (spectra 3 and 4) attributed to S2
(250 nm), S2 (295 nm), S2O42, c-S8 (320 nm) and S22 (385 nm) species. They were
largely removed after evacuation at 573 K or transformed into chain molecules
398 H. Frster
Fig. 34. Transmission UV-VIS spectra of reactive interaction of H2S and SO2 over zeolite NaX.
After adsorption of H2S at room temperature (1); after treatment in the IR beam for 1.5 h (2),
giving rise to the IR spectrum A (see inset); after admission of SO2 at room temperature (3);
after treatment in the IR beam for 1 h (4) resulting in the IR spectrum A (see inset); after evac-
uation for 3 h at 573 K. Reprinted from [100] with permission of Academic Press, Inc
like S3 and S4 (spectrum 5). On the other hand, S2 species proved to be inactive
in the Claus reaction. In conclusion, combined transmission UV-VIS and IR spec-
troscopy is a suitable tool for the study of the details of the Claus reaction over
zeolite catalysts.
4.2.2
Organic Compounds
For simple organic molecules with a single chromophore there are a number of
possibilities for electronic excitation. Saturated compounds with single bonds
allow only ss* transitions occurring in the far UV region close to the Rydberg
transitions. In the case of lone-pair electrons, e.g., on heteroatoms, the generally
weaker ns* and np* transitions are possible at lower wavelengths. At last
p-bonding in chromophores gives rise to pp* transitions. Benzene, e.g.,
exhibits three bands around 180 (A1gE1u, b), 200 (A1gB1u, p) and 260 nm
(A1gB2u, a) that in the nomenclature of Clar are denoted as b, p and a bands
[3]. Hckel MO treatment yields six wave functions, as shown in Fig. 35 [6, 9].
Their symmetries, given in brackets, are irreducible representations of point
group D6h. The ground electronic state is given by (a2u)2(e1g)2(e1g)2 which belongs
to the totally symmetric species 1A1g. The lowest excited state has the electronic
structure (a2u)2(e1g)2(e1g)(e2u). According to group theory, the symmetry species
of the orbital part of this electronic wave function are obtained from the direct
product G=e1ge2u=B2u+B1u+E1u. Taking electron repulsion into account the term
Fig. 35. The p molecular orbitals according to Hckel MO calculations (left), energy level
diagram (center) and term diagram of benzene with electronic transitions indicated
(right). Quoted from [6]. Copyright John Wiley and Sons, Ltd. Reproduced with permission.
The other parts are taken from Chang R (1971) Basic principles of spectroscopy. McGraw-Hill
Kogakushi, Ltd, Tokyo. Reproduced with permission of the McGraw-Hill Companies
level diagram is as shown on the right-hand side of Fig. 35 with the transitions
indicated by arrows. The a band, although symmetry-forbidden, appears with a
vibrational fine structure due to the distortion of the benzene ring by a bending
vibration which allows it to borrow intensity from the b band [3]. Indeed the 00
vibrational transition is not observed which becomes allowed upon symmetry
lowering by substitution of the aromatic ring or by interaction with solid surfaces
[103]. Early investigations on aromatic systems adsorbed on zeolites have been
carried out by Russian workers, e.g., [104, 105]. Some of the most important
results have been reviewed [16, 106]. Benzene shows molecular adsorption with
small band shifts to higher wavelengths. The larger displacements in the case of
amine derivatives are explained by respective changes in the levels responsible
for the np* transitions as the result of participation of the lone-pair electrons
of the substituent in the interaction. Comparing the frequency shifts of different
alkali metal ion-exchanged zeolites X this shift could be correlated with the elec-
trostatic potential of the nonframework cation expressed by the charge/radius
ratio, as shown in Fig. 36. Triphenylcarbinol and anthraquinone on NaY proved
only molecular adsorption [16]. The same is true with benzene, toluene and
cumene on CaX. However, with aniline, diphenylamine, triphenyl methane and
triphenylcarbinol, new bands appeared which must be assigned to carbocations.
Aromatic systems have always attracted considerable attention. Eremenko
et al. [107, 108], choosing naphthalene and naphthylamine as probe molecules in
alkali metal and alkaline earth ion-exchanged faujasites, investigated the forma-
tion of donor-acceptor complexes and the oxidation of naphthalene using
luminescence and DR spectra. Naphthalene adsorbed from hexane solution in
400 H. Frster
Fig. 36. Wavelength shift Dl of the electronic band of aniline (1), p-phenylenediamine (2), pyri-
dine (3) and nitrobenzene (4) adsorbed on zeolites LiX, NaX, KX, RbX and CsX as a function
of the electrostatic potential U=e/r () of the respective nonframework cation. Reprinted with
permission from [106]. Copyright 1971 American Chemical Society
zeolite NaX exhibits both fluorescence (310350 nm) and phosphorescence

(470540 nm) indicating monomer physisorption on different sites in the zeolite
cages.An additional intermediate unstructured band with a maximum at 400 nm
has been attributed to a naphthalene excimer, i.e., an excited dimer. Loading the
zeolite from the vapor phase resulted in a broad unresolved luminescence of CT
complexes and dimers on the outer zeolite surface. Coadsorption of water pro-
motes the splitting of the dimers and migration into the zeolite interior.Adsorbed
oxygen in the presence of water vapor caused oxidation of naphthalene to a-
naphthoquinone in the case of sodium ions and to an a-naphthoquinone/b-
naphthol mixture in the case of alkaline earth ions. Similar behavior was
observed for naphthalene in zeolite NaY where as well as weakly bound
monomers chemisorbed donor-acceptor CT complexes but no exciplexes or
excimers were evidenced [108]. As electron-acceptor sites coordinatively unsat-
urated silicon and aluminium ions are suggested.
These last considerations take this topic into the realm of electron- and
charge-transfer reactions; of those a vast number have been investigated in zeo-
lites, recently reviewed by Yoon [109]. The reader is referred to this article for fur-
ther information.
4.2.3
Reactive Intermediates: Carbocations, Carbanions, Radicals
In the discussion of reaction mechanisms of multi-step reactions the so-called

reactive intermediates like carbocations, radicals, carbanions and carbenes
play an essential role as they function in a way as relay stations in consecutive
reactions. Due to their mostly short lifetimes and/or low concentrations their
presence is easier postulated than proved. Concerning carbocations and review-
ing the hitherto obtained results by means of optical spectroscopy, electronic
spectroscopy seems to be superior to vibrational spectroscopy due to (i) the
considerably larger absorptivity and (ii) the lack of obscuring adjacent bands,
provided there is not much interference with bands of the host, e.g., with d-d
transitions in the case of TMI-exchanged zeolites or with CT bands. Although
the electronic spectrum basically contains the overall information on the rota-
tional, vibrational and electronic degrees of molecular freedom, those of the two
former cannot be readily extracted, as the rovibrational fine structure frequently
remains unresolved. Therefore, it must be considered as a substantial drawback
that no structural evidence is given by electronic spectroscopy so that additional
vibrational spectroscopic studies are necessary resulting in a mutual support.
On the zeolite surface the Brnsted acid sites are responsible for the genera-
tion of (i) carbonium ions from paraffins, (ii) alkyl carbenium ions from olefins,
and (iii) alkenyl carbenium ions from dienes by protonation, while on Lewis acid
sites (i) alkyl carbenium ions from paraffins and (ii) alkenyl carbenium ions from
olefins are formed by hydride ion abstraction.
In carbocations formed on solid surfaces chromophores must be present with
transitions in the near UV or VIS region, i.e., in a spectral range accessible to
commercial spectrometers. This excludes alkyl carbocations from spectral proof
because they have no absorptions above 210 nm [110]. Unsaturated compounds
show p-p* transitions in the far UV region which upon insertion of alkyl groups
undergo a bathochromic shift due to hyperconjugation. Conjugation of olefinic
double bonds gives rise to a lowering of the p orbitals accompanied by a reduc-
tion of the HOMO-LUMO gap resulting in a red shift together with an increase
in intensity. This forms the basis for the detection of alkenyl carbenium ions by
UV-VIS spectroscopy.
Early IR and UV-VIS spectroscopic studies on the formation of carbonium
ions from triphenyl methyl compounds on zeolites, titania and alumina were car-
ried out by Karge [111]. In 1979, upon interaction of olefins like ethene and
propene with zeolites CoNaY, NiCaNaY, PdNaY and HY, the appearance of elec-
tronic bands between 230 and 700 nm was observed by Garbowski and Praliaud
and attributed to an allylic carbenium ion which upon thermal treatment trans-
forms into polyenyl carbenium ions and/or aromatic compounds [112]. These
findings were corroborated and extended by studies of the interaction of
propene, cyclopropane and trans-butene on zeolites NaCoY and HM [30]. In spite
of the obscuration of the spectrum in the range between 450 and 700 nm by the
threefold split d-d band of tetrahedrally coordinated Co(II) ions in the case of
zeolite NaCoY, the development of bands near 330, 385 and 415 nm was assigned
to unsaturated carbocations.
402 H. Frster
Systematic transmission spectroscopic studies of the transformations were

carried out on the systems allene, propene, cyclopropane, isopropanol and ace-
tone over NaM, HM and NaHM [113117], 1-butene over NaHY [117, 118], allene,
propene and cyclopropene over NaHY and NaHZSM-5 [116, 117, 119122], neo-
pentane over NaY, NaHY, NaHM and NaHZSM-5 [119, 123, 124], cyclohexene over
HZSM-5 [119], 1-hexene, cyclohexane, cyclohexene, cyclohexadiene and benzene
over NaHY and NaHZSM-5 [117, 125], allyl alcohol, acrolein and allene over
HZSM-5 [126], allene and propyne over nonacidic zeolites NaA, CoNaA, CoNaX,
NaY and NaHY [121] and benzene, toluene, ethylbenzene, and cumene over
HZSM-5 [127].
The spectral features observed are in general quite similar to those shown
in Fig. 37 with zeolite NaHY exposed to propene as an example [119]: Shortly
after adsorbate admission a band around 300 nm emerges, increasing in intensity
and shifting to higher wavelengths with time (a-d) accompanied by the devel-
opment of bands around 370 and 450 nm enhanced by thermal treatment. Upon
evacuation at higher temperatures, lmax is displaced to shorter wavelengths (e, f).
This triad of bands is attributed to the p-p* transitions of mono-, di- and tri-
alkenyl carbenium ions. Basically, alkenyl carbenium ions can form by (i)
protonation of dienes, (ii) hydride ion abstraction from olefins on Lewis acid
Fig. 37. Transmission electronic spectra of zeolite NaHY loaded with 1.33 kPa propene at room
temperature after 5 min (a), 1 h (b), 2 h (c); at 320 K after 1 h (d); after evacuation for 1 h at
room temperature (e), at 400 K (f ) and at 570 K (g). Reprinted from [119] with kind permis-
sion of Elsevier Science NL, Sara Burgerhartstraat 25, 1055 KV Amsterdam, The Netherlands
sites, (iii) intermolecular hydride ion abstraction from olefins by alkyl carbenium
ions, (iv) conjunct oligomerization, or (v) cracking of oligomers [128]. Mono-,
di- and trialkenyl carbenium ions must contain at least 3, 5 or 7 C atoms, respec-
tively.
The most important results from electronic spectra are briefly summarized as
follows:
The capability for alkenyl ion formation from different hydrocarbons and their
derivatives follows the same sequence as observed in superacid solutions:
dienes>olefins>alcohols>paraffins. Decreasing the Brnsted and the overall
acidity has the consequence of slowing down the formation of alkenyl carbe-
nium ions. Thus, protonation and hydride ion abstraction are the triggering
steps.
The wavelengths of the band maximum lmax characteristic for alkenyl carbe-
nium ions formed in zeolites agree approximately with those calculated from
the Sorensen equation
lmax = (319.5 + 76.5 n) (37)
describing in superacid solutions the frequency dependence of polyenylic car-

benium ions of the type
Me2C-CH-[CH-CH]n-C+Me2
as a function of the number of conjugated double bonds.

lmax is sensitive to structural changes of the carbenium ion; e.g.:
1. insertion of an additional double bond gives rise to a wavelength jump of
ca. 60 nm, while
2. extension of the carbon chain or formation of branched oligomers results
in a smooth upscale wavelength shift accomplished by consecutive reac-
tions.
Therefore, only information on the degree of unsaturation, i.e., number of
double bonds, is available while the structure remains uncertain.
Starting, e.g., from propene (as shown in Fig. 37), the consecutive reaction
with itself gives rise to an equilibrium mixture of various monoenyl carbenium
ions the common feature of which is a C6 backbone. This explains the different
wavelengths or the broad wavelength ranges observed in the spectra. In-
creasing the temperature enhances further oligomerization via alkyl and alkenyl
cations. Upon protonation cyclic hydrocarbons undergo ring opening to olefins.
Following the temporal development of the peak heights the induction period
observed with dienyl ions points to their consecutive formation from monoenyl
ions.
The higher the amount of Brnsted acid sites, the more pronounced the
oligomerization of olefins and, consequently, the formation of alkenyl carbenium
ions absorbing at higher wavelengths. It is assumed that a decrease in the num-
ber of Brnsted acid sites is related to an increase in the acid strength, thus pro-
moting the stability of the respective carbenium ion.
404 H. Frster
The explanation of the upscale frequency shift linked with increasing tem-
perature might be cracking and/or desorption, so that the remaining ions are of
shorter carbon chain type, thus absorbing at higher frequencies.
Complementary aid is given by combined IR experiments from which bands
at 1535 and 1505 cm1 have to be assigned to the C-C-C stretching modes of
mono- and dienylic carbenium ions, respectively.At low temperatures there is no
indication of aromatics.After evacuation sometimes a band at 1585 cm1 remains
which must be attributed to coke.
Quantum chemical methods are valuable tools for assisting spectroscopic
studies. Semiempirically calculated lowest vertical singlet excitation energies
(LVSEE) prove the red shift of the bands with growing carbon chain length and
methyl substitution [122]. Increased methyl substitution increases the stability of
the ion as shown from the standard heat of formation DHf. DHf on the other hand
slightly decreases in the sequence mono->di->trienyl carbenium ions. Their ease
of formation as well as stability depends on zeolite acidity, i.e., on their environ-
ment that may be approximated by the relative permittivity from an MNDO
effective charge model correlated with the heat of solvation, DHsolv, paralleling the
stability of the alkenyl carbenium ion. From quantum chemical studies also a hint
to the site of H+ attack and thus to the kind of carbocation intermediate can be
given, as shown for the protonation of allyl alcohol. The results of the quantum
chemical calculations prove once more that the assignment of a given UV-VIS
band to a well-defined alkenyl carbenium ion is rather difficult; only a group of
such ions may be indicated.
There are a number of processes for which a carbenium ion mechanism has
been proposed, and UV-VIS spectroscopy has proved to be a very attractive tech-
nique in these cases. However, interpretation of the spectra is rather difficult due
to the overlapping bands from simultaneous formation of several carbocations
and the shift of the peak position effected by the adsorbent, so that comparison
with the bands observed in superacid solution is hampered. This technique was
used to explore the formation of aromatics from methanol over zeolite HZSM-5
[129]. Cyclization of dienyl to cyclopentenyl cations (absorbing at 280 nm) and
their structural rearrangement into cyclohexenyl carbenium ions (~300 nm) as
well as further generation of polyalkylaromatics and polyaromatics (260, 280,
330, 420, 500, 600 nm) were observed.Although no coking in the zeolite channels
during the experiment was detected, polyalkylaromatics and polyaromatics are
considered to be the precursors for coking which obviously in this case takes
place on the outer zeolite surface.
An interesting variant of the spectroscopic experiment is in situ investigations
under catalytic conditions. For this purpose a special reactor cell has been
designed equipped with quartz windows which allows the monitoring of the
course of the reaction on the catalyst wafer. As an example, the transformation
of ethene over HZSM-5 was studied up to 573 K using a scattered transmission
accessory in which the sample is placed directly in front of the multiplier, so that
the forward-scattered light almost completely impinges the detector [130]. At
lower temperatures no substantial differences could be established between the
carbocation spectra obtained in situ and ex situ. The spectroscopic discrepancies
in the peak positions above 548 K are tentatively explained by the different
stability of carbocations at room temperature compared with higher tempera-

tures.
Coke represents undesired carbonaceous residues leading to catalyst deacti-
vation necessitating a burning off of used catalysts, e.g., in the FCC process. Karge
et al. [29, 131] carried out systematic studies on coke formation in zeolites. Trans-
mission electronic spectroscopy enabled these authors to distinguish between
different stages of coke formation, low- and high-temperature coke, as can be
seen from Fig. 38 in which the difference spectra of a methanol-loaded ZSM-5
catalyst are displayed obtained upon subsequent heating to the indicated
temperatures in a closed cell. While at temperatures lower than 550 K essentially
bands of mono- and dienylic carbocations near 300 and 360370 nm appear,
the high-temperature range is dominated by absorptions around 430 and
above 500 nm of small aromatic carbocations and polyaromatics, respectively.
The nature of the different forms of coke thus formed turned out to be dependent
not only on the temperature but also on the acidity, morphology and crystallite
size of the zeolites. Generation, characterization and transformations of un-
saturated carbenium ions in zeolites have been comprehensively reviewed
recently [132].
Carbanions as intermediates have been suggested in the base-catalyzed iso-
merization of unsaturated nitriles.As long as zeolites with anion-exchange prop-
erties are not available, basic centers must be introduced by exchange of the non-
framework ions by alkali metal ions or of silicon by germanium, by precipitation
Fig. 38. UV-VIS difference spectra of carbonaceous deposits after reaction of methanol on
HZSM-5 obtained from thermal treatment of the system at the indicated temperatures.
Reprinted from [131] with kind permission of Elsevier Science NL, Sara Burgerhartstraat 25,
1055 KV Amsterdam, The Netherlands
406 H. Frster
of basic oxides within the zeolite interior or by generation of ionic alkali metal
clusters, e.g., by impregnation with sodium azide followed by controlled decom-
position. Applying the latter procedure to NaY, a modified basic faujasite was
obtained which, exposed to allyl cyanide, readily catalyzed its isomerization to
crotononitrile, proven by IR spectroscopy [133]. In the literature the electronic
bands of carbanions are reported to appear in the 350550 nm range. The UV-
VIS spectra recorded during transformation showed only one broad band
around 400 nm that increased in intensity and shifted to higher wavelengths with
increasing time of contact and/or temperature. As it did not show the character-
istic spectral behavior reported for unsaturated carbenium ions, it is assumed to
be due to a p-p* transition of the carbanion formed by interaction of allyl
cyanide with basic sites.
Radicals are very reactive and therefore short-lived species with one or more
unpaired electrons and may be neutral or charged as radical cations and anions.
Neutral radicals often result from homolytic cleavage of a covalent bond to which
the necessary dissociation energy is frequently supplied by heat or radiation.
Therefore, early spectroscopic studies of radical formation on zeolites were car-
ried out with molecules adsorbed in zeolites under UV irradiation; these have
already been comprehensively reviewed in the literature [15, 16, 105, 106]. Ben-
zene, e.g., in zeolite NaX, did not reveal spectral changes upon irradiation, while
adsorbed in zeolites CaX and HX two bands at 460 and 565 nm were formed
from which the former was assigned to the cyclohexadienyl cation and the
latter to the radical cation C6H6+. Similar results were obtained with cumene and
anthraquinone. There are many studies of free-radical reactions in zeolites, in
particular upon photolysis of molecules yielding fragments which react accord-
ing to the cage effect [134]. Photoexcitation of the CT band of donor-acceptor
precursor complexes by lasers leads to the formation of transient CT ion pairs
composed of radical ion pairs which are stabilized in the zeolite cages by retar-
dation of the back electron transfer and can be proved by UV-VIS spectroscopy
[109]. Also, radical formation without irradiation is reported in the recent liter-
ature. Loading zeolite NaY and its protonated forms with bi- and terthiophene
gave rise to bands in the 400800 nm range attributed to bi-, ter- and hexathio-
phene radical cations [135].
4.2.4
Polymerization in Zeolite Channels
Zeolites provide novel hosts for preparation of conductive polymeric materials

within their channels [136]. Provided the synthesis and the alignment of conju-
gated polymers are carefully controlled the formation of oriented molecular
wires for microelectronic applications such as signal processing and information
storage is feasible. As already outlined in the last section, acid groups in the zeo-
lite are responsible for the generation of thiophene oligomers [135].
As another example, the protonated forms of zeolites M, omega, L, Y, beta,
ZSM-5 and SAPO-5 have served as media for the alignment and packaging of
incarcerated methylacetylene in order to explore the concepts for the production
of tailored polymers with possible useful nonlinear optical properties [137].
Upon irreversible uptake the formation of yellow to brown products was the
result of initial protonation to produce a vinyl cation followed by oligomeriza-
tion. Several broad bands between 300600 nm tailing out to 750850 nm were
observed, as shown for methylacetylene/HZSM-5 in spectrum b of Fig. 39, and
these were assigned to conjugated oligomers pointing to similar mixtures of
products in all zeolites under investigation. The conjugated species seemed to be
a small portion of the total, as the spectrum of longer conjugated chains would
be more extended to the long-wavelength range.
Oxidative doping enhances conductivity in conjugated polymers. Using iodine
as oxidant in the case of methylacetylene/HZSM-5 gave UV-VIS spectra that
essentially look like a superposition of the separate spectra of I2 and methy-
lacetylene in HZSM-5 (Fig. 39) and confirm that, as electron-deficient entities, the
cationic species will not be readily oxidized further.
A more detailed description of the extent of branching and cyclization is not
available from the spectra, as no obvious differences are obtained even using zeo-
lites with different spatial restraints.
An interesting variant of the formation of nanometer-sized conductive strands
is pyrolysis of intrazeolite polymers such as polyacrylonitrile in zeolite NaY, the
color of which turns from white to grey-black [136]. The electronic spectrum of
the extracted material is practically structureless with a feature at about 350 nm
and resembles that of graphite. In this way graphite-like chains with conductiv-
ity in the order of 105 S cm1 can be prepared.
Fig. 39. Diffuse reflectance UV-VIS spectra of iodine in HZSM-5 (a), methylacetylene in
HZSM-5 (b) and coadsorbed iodine and methylacetylene in HZSM-5 (c). Reprinted with per-
mission from [137]. Copyright 1991 American Chemical Society
408 H. Frster
4.2.5
Clusters of Metals, Metal Ions and Semiconductors in Zeolite Hosts
Currently, there is a trend towards ever-decreasing electronic devices [138, 139].

The reduction in size of a solid leads to clusters with novel properties which have
to be placed somewhere between those of the bulk phase and those of molecules
or atoms. Changes in color are indicative for such quantum size effects. For these
small clusters, with respect to their size also called nanoclusters, the description
of the electron motion in a periodic structure by energy bands becomes replaced
by the image of discrete energy levels similar to the particle-in-a-box behavior.
One of the approaches to produce highly dispersed materials in a well-defined
way is their preparation within the porous structure of zeolites that are known
to stabilize unusual molecular species. Candidates of potential application of
such nanoscale particles might be composite superconductors, new magnetic
materials, effective catalysts in the case of metal and metal ion clusters and light
switches, photocatalysts, reversible image capture/storage devices, solid-state bat-
teries, electrochromic devices, sensors and solar cells in the case of semicon-
ductor clusters. The preparation occurs either (i) by direct incorporation from
the melt or the vapor or (ii) by in situ creation such as ion exchange followed by
chemical reactions (reduction, sulfidation), thermal decomposition of inorganic
as well as organometallic precursors or g-irradiation of zeolites. Electronic spec-
troscopy is also a sensitive tool for the characterization of the electronic struc-
ture of these clusters.
The state of the art of silver clusters in zeolites, their structure and chemistry,
complex formation and application have been reviewed by Sun and Seff [140].
Water cleavage is among others one of the most desirable processes for storage
of solar energy. Upon illumination with light near 400 nm of silver-loaded zeo-
lites in aqueous dispersion,Ag+ ions become reduced by simultaneous evolution
of oxygen [141]. This process is accompanied by self-sensitization, i.e., in the
course of this reaction the sensitivity bathochromically shifts from the near UV
to the visible region. This process can be rationalized as LMCT transition from
the oxygen lone pair to the 5s level of zeolite-embedded Ag+ ions. The lifetime of
the holes hopping between the zeolite oxygens is long enough for their accumu-
lation and reaction with water. The self-sensitization can be explained by the for-
mation of partially reduced silver clusters absorbing at lower energies. Coupling
of the oxidative step to a transition-metal complex forces the reduction of water
under evolution of hydrogen via a MLCT transition. There is still little informa-
tion on the details of the silver species produced which seem to be stronger
reductants compared to bulk silver. A study on silver cluster stabilization in zeo-
lites of different pore sizes and composition while also estimating the cluster size
by comparison of the results from DRS, electron microscopy and small-angle X-
ray scattering was carried out by Bogdanchikova et al. [142].
In contrast to zeolites X,Y and erionite, the amount of clusters formed in mor-
denite was rather large, i.e., the cage structure of the latter proves to be ideal for
silver cluster stabilization. In the electronic spectra of AgM upon reduction with
hydrogen at different temperatures (Fig. 40), two intense bands at 320 and 285 nm
dominate beside two weak bands at 410 and 235 nm as well as shoulders at 350,
Fig. 40. Diffuse reflectance UV-VIS spectra of silver mordenite (Si/Al=15) reduced by hydro-
gen at room temperature (1), 323 K (2), 373 K (3), 473 K (4), 573 K (5) and 773 K (6). Reprinted
from [142] with kind permission of Elsevier Science NL, Sara Burgerhartstraat 25, 1055 KV
Amsterdam, The Netherlands
310 and 260 nm. The assignment takes its cue from that of Henglein et al. [143]
for silver particles in aqueous solution, although some caution is advised as the
band position depends on the environment and differs somewhat, e.g., for noble-
gas matrices, liquid and frozen solutions and zeolites. Larger colloidal silver par-
ticles are discernible by the narrow surface plasmon band of the electron gas at
380 nm indicating metallic properties. This band becomes broadened and dis-
placed to 360 nm in the case of quasi-metallic particles with agglomeration num-
bers around 10. Sharp bands below 360 nm are attributed to nonmetallic
oligomeric species, called clusters. From smaller species Ag0 absorbs at 360 nm,
Ag2+ at 310 nm and Ag42+ at 275 nm. With respect to AgM the band at 410 nm cor-
responds to metallic silver, the 320 nm absorption to silver clusters of 0.7 nm size
and those at 310, 285, and 235 nm to Ag2+, Agn+ and Ag+ species, respectively.
Reduction at 773 K shows agglomeration of the 320 nm clusters to larger parti-
cles absorbing at 360 nm, while the Ag2+ dimers at 310 nm are stabilized and not
reduced even at this temperature.
DRS has also been successfully applied to follow the kinetics of silver-cluster
formation in AgX monitoring the intensity changes of the 410 nm band which is
assumed to originate from Ag32+ clusters in the SI-SI-SI position of the hexag-
onal prisms [144]; The rate-determining step has been rationalized by a model
of diffusion into a sphere.
So far in addition to uncharged metal clusters, charged species have also been
present. Prototypes of the latter are the alkali metal ion clusters which can be
formed in zeolites by exposure to high-energy radiation, by alkali metal doping
or by introduction of molecular electron donors [109, 134]. These seemingly very
different procedures can be brought down to the common denominator of dis-
410 H. Frster
solution of surplus electrons in traps formed by the counterions of the zeolite

framework.As mentioned earlier, the zeolite takes the role of a polar solid solvent
and the widespread case of alkali metal doping, e.g., in faujasites, may be
described as trapping of the released electron from the incoming alkali metal
atom M among four equivalent sodium cations
M + 4 Na+zeolite M+ + (Na+)4 e (38)
In this way paramagnetic clusters of the type Mn(n1)+ with n=26 have been
observed with sodium, although from MAS NMR and magnetic studies evidence
on diamagnetic clusters Mnm+ with n+m even has been reported. The main
technique applied for identification of the cluster size is ESR. From the number
of hyperfine lines predicted by the quantity (2nI+1), where I is the nuclear
spin, the number n of equivalent nuclei M+ in the cluster can be evaluated.
The most frequently observed species in the case of sodium are Na43+ percepti-
ble from their 13 hyperfine line patterns. More complex ESR spectra with in-
tensity ratios deviating from binomial distribution have sometimes been
observed indicative of coexisting clusters of different nuclearity. For a more
detailed discussion, which is beyond the scope of this article, the reader is
referred to several reviews that represent the present state of alkali metal cluster
research [140, 145148].
The technique most frequently used next to ESR for the characterization of
alkali metal ion clusters has been electronic spectroscopy. This seems under-
standable, as the trapped electron forms a color center quite similar to the F-cen-
ters in alkali halides (see Sect. 2.4.4). Both can be generated by exposure to g-rays
or UV light and can be bleached upon irradiation by intense white light, a pho-
tochromic behavior which gives rise to application as a read/write device. Unfor-
tunately, the bands are broad and featureless without vibronic structure. As an
example, the DRS of the sodium-doped sodium sodalite in Fig. 41 are dominated
by a transition between the Stark effect broadened ground and first excited elec-
tronic states of the cluster in the range 15,50018,000 cm1 accompanied by an
UV band near 38,000 cm1. There have been several attempts to simulate the spec-
trum of the solvated electron in sodalite for a fixed and averaged orientation as
shown on the right-hand side of Fig. 41 in comparison with the experimental
spectrum at the lowest loading Na43+/Na33+=1/50 indicating reasonable agreement
[145]. For another approach to the simulation of the electronic spectrum of the
Na43+ color center in sodalite using 3-d wavepacket calculations, the reader is
referred to an article by Blake and Stucky [147] and the references cited therein.
These calculations reproduce the spectra well, support the observed linear rela-
tionship between lmax of the guest band and the lattice constant of the host, and
are in accordance with the distance of the color center to the nearest Na+ neigh-
bors deduced from ESR experiments. In the a1 ground state the solvated electron
is delocalized near the center of the (Na+)4 tetrahedron, while the excited state is
a triply degenerate t2 state. Upon excitation the resultant deshielding causes a
vibrant breathing of the cluster that, apart from other effects, is jointly respon-
sible for the breadth of the absorption band evidencing the sensitivity of the
band to the Na+-Na+ distance in the cluster.
UV/VIS Spectroscopy
Fig. 41. Electronic spectra of Na43+ centers in sodium sodalite. Left Experimental spectra at Na43+/Na33+ loadings of 1:2, 4, 10 and 50. Right Calculated
single-orientation spectrum and orientationally averaged spectra of a Na43+ cluster in a sodalite cage at high dilution with atomic charges of Si=+1.5,
411
Al=0.85, O=0.84 and Na=+1 (top). Blown-up experimental Na43+ spectrum at Na43+/Na33+=1/50 for comparison (bottom). Reprinted with permission from
[145]. Copyright 1992 American Chemical Society
412 H. Frster
Attempting to summarize the hitherto existing literature on electronic spec-

troscopy of alkali metal ion clusters in zeolites [145147, 149168], it can be
stated that the absorption of Na43+ clusters in zeolite NaY is correspondingly
observed around 500 nm, while for X zeolites bands between 500 and 550 nm
appear according to the method of preparation and the Si/Al ratio. Both absorp-
tions were also proved for the first time after decomposition of sodium azide in
faujasites [161, 162]. These bands are assigned to the 1s1p transition of the
electron trapped in the cluster. Contradictory are the data for Na43+ in sodalite
with 560 nm [150] and 628 nm [146, 159] presumably due to the different type of
synthesis. In zeolite NaA at low loading two bands are detected at 400 and 443 nm
attributed to single-electron excitation of clusters containing one and more
captured electrons, respectively, while the bands near 590, 886 and 1033 nm upon
higher loading are assigned to an electron-hole pair formation and a surface
plasmon transition, respectively.
At high amounts of doping the color of the sample turns black accompanied
by a broadening of the absorption to a continuum with an onset at 1550 nm
(0.8 eV). From these findings it must be concluded that the material has now
become a semiconductor with a band gap of 0.8 eV. At this high level of alkali
metal uptake an interaction of neigboring Na43+ clusters by overlap of their wave
functions and coupling through exchange forces has to be assumed which results
in the collapse of the ESR hyperfine structure to a singlet and raises hopes for
achieving the insulator-metal transition at high excess electron concentrations,
i.e., for the preparation of metallic conducting zeolites.
Light switching can be performed with semiconductors at their band edge by
laser illumination which, due to the third-order non-linear optical properties,
changes the refractive index and thus their state from opaque to translucent. As
extreme speeds for processing are desired, the switching frequency must be
increased by removal of the free carriers into defect (i.e., surface) sites which is
simply achieved by increasing the surface/volume ratio, i.e., decreasing the par-
ticle size of the semiconductor. Making use of the zeolite pore structure, a num-
ber of different nanosized elemental and compound semiconductors have been
prepared [139]. As an example of the latter, CdS clusters were generated by Cd2+
ion exchange followed by sulfidation. In zeolite Y distorted Cd4S4 cubes were
formed in the sodalite unit where the Cd ions were octahedrally coordinated to
three S and three O atoms of the zeolite 6R window [163, 164]. While bulk CdS is
yellow-orange colored with a band gap of 2.42 eV (512 nm) [13] absorbing in the
blue part of the visible range and increasing in absorption towards the ultravio-
let, in zeolite Y at low concentrations the spectrum of the discrete cubes is dis-
placed downscale with a shoulder at 280 nm corresponding to an edge shift of
230 nm. At a threshold of about 4 wt.% CdS, the population of adjacent sodalite
cages becomes possible, and with increasing loading the spectral pattern shifts
to the red with a distinct shoulder near 350 nm (see right-hand side of Fig. 42)
which might be due to the exciton transition. Further increase of CdS con-
centration gives rise to a progression towards a semiconductor hyperlattice,
as shown on the left-hand side of Fig. 42, with an intermediate behavior between
the discrete cubes and the bulk. While at low amounts no emission was ob-
served, above the threshold the interconnection brings about luminescence,
UV/VIS Spectroscopy
Fig. 42. Left Ultimate hyperlattice arrangement of (CdS,O)4 cubes in adjacent sodalite cages of zeolite Y (hatched=Cd, open=S, O omitted for clarity). Right
Absorption (dotted), excitation (solid) and emission (dashed) spectra of 6.5 wt.% CdS occluded in zeolite Y. From [139] with kind permission from Kluwer
413
Academic Publishers
414 H. Frster
as shown on the right-hand side of Fig. 42, with an excitation spectrum centered
at the exciton shoulder. The properties of the semiconductor can be varied by appli-
cation of different hosts.As revealed by electronic spectroscopy, the cluster-cluster
interaction in CdS-loaded zeolite A is much weaker than in zeolite Y [163, 164].
4.2.6
Zeolite-Hosted Oxides
Synthesis of molecular sieves with T sites isomorphously substituted by other

elements is frequently accompanied by the formation of more or less oxidic
extraframework species as by-products (see Sect. 4.1.2.2). In this section the
intended preparation of finely divided oxides or oxide clusters within the voids
of the framework will be discussed.
Zeolite-stabilized oxides are of interest for catalytical application, due to their
redox behavior as potential sensor materials and owing to the revealed quantum
size effects for highly dispersed semiconductors.
Preparation of zeolite-hosted oxides can be carried out either (i) by chemical
vapor deposition (CVD), i.e., loading the zeolite with volatile chlorides like TiCl4,
VOCl3 or SnCl4 in a nitrogen carrier stream, followed by hydrolysis and calcina-
tion, or (ii) by impregnation with suitable compounds such as SnCl4 or VO(acac)2
in aqueous or organic solution and subsequent calcination.
Due to their different appearance as
isomorphously substituted ions in T positions,
molecular oxidic species attached to the zeolite framework,
quantum-sized particles, and
oxides with bulk properties
there is considerable need for their reliable identification. This can be achieved
from the maximum or the onset of absorption, i.e., the intersection of the tangent
through the inflection point with the abscissa, or from both of the UV-VIS DR
spectra. In the case of TiOx-loaded faujasites it was possible to distinguish
between TiOx species double- or single-bonded to the framework and bulk
TiO2 (anatase) by their different maximum/onset at 240 nm/300310 nm,
290 nm/330340 nm and 328 nm/370 nm as well as the proof of tetrahedral Ti4+
from the onset at 250260 nm [165].
Vanadium oxide particles in HZSM-5 prepared by CVD displayed CT bands at
210 and 250 nm typical of V in a distorted tetrahedral environment and one at
375 nm attributable to V in a distorted octahedral environment pointing to
extraframework VOx species enriched at the external surface according to XPS
results [166]. Reversible reduction and oxidation was monitored by the decreas-
ing or increasing reflectance at fixed wavelength (620 nm), due to the increasing
or decreasing number of possible d-d transitions. Compared to VS-1,VZSM-5 is
more sensitive to reduction/oxidation and shows a faster response to oxygen than
to hydrogen.
Vanadium has also been incorporated into mesoporous MCM-41 materials
by hydrothermal synthesis, impregnation and CVD [167]. UV-VIS DR spectra
reveal predominantly mononuclear vanadium(V) oxide and oligomeric clusters
Fig. 43. Left Reversible reduction/oxidation cycles upon exposure to CO and O2 of SnNaY
loaded by impregnation (A) and bulk SnO2 (B). Right Detection of small amounts of CO in air
(A 250 ppm; B 500 ppm; C 1000 ppm; D 2000 ppm) by the decrease of the reflectance of SnNaY
at 16200 cm1. Reprinted from [168] with kind permission of the Materials Research Society,
Warrendale
both in more-or-less tetrahedral coordination inside the pore system. Both can
be reduced by hydrogen, the latter with a steeper response. Irreducible vanadium
is either buried in the framework or in nonporous by-products.
SnOx particles in Y zeolites have been formed with diameters <1 nm using
CVD and with diameters of 34 nm by impregnation indicating quantum size
effects from the blue shift of the absorption onset in comparison to bulk SnO2
[168]. While reduction of SnO2 with hydrogen is irreversible, it becomes
reversible with CO, leading to the formation of SnOx particles (1<x<2). As the
nanoparticles are located in the pores near the external zeolite surface they are
highly accessible to gases. Figure 43 (left) shows reversible reduction/oxidation
cycles of SnNaY prepared by impregnation compared to bulk SnO2, monitored by
the change of reflectance at fixed wavenumber (16,200 cm1) and demonstrate
their application as sensors with optical detection. The detection of low amounts
of CO down to 250 ppm is proved on the right-hand side of Fig. 43.
4.2.7
Zeolite-Encapsulated Carbonyl, Metal Chelate Complexes and Dyes
In this section, guests confined within the zeolite cavities will be discussed, i.e.,
guests of sizes larger than the portals formed by the pore aperture, so that they
are unable to escape. Steric restrictions and the strong intrazeolitic electric fields
affect the chemistry in these nanoreaction chambers and may cause structures
and properties that are modified compared to those obtained in solution or on
external surfaces. Furthermore, entrapping hinders interaction between the clus-
ters and enhances their stability.
Different strategies for encapsulation of bulky entities in zeolite hollows have
to be distinguished [79, 169]:
1. in situ preparation from the components,
2. flexible ligand method,
416 H. Frster
3. template synthesis method, and

4. zeolite synthesis method.
The topic of clusters confined to zeolite cages in general and of organometallic

clusters in particular has been reviewed by several authors, e.g., [48, 79, 169171].
In these articles the methods of synthesis have been described in detail, so that
we can restrict ourselves to outlining their characterization which has profited
greatly from electronic spectroscopy.
Prior to characterization encapsulation must be ensured and clusters formed
outside the cavities must be ruled out. Only then can characterization be reliably
carried out. A battery of techniques is available for this purpose, such as 13C, Xe
and metal NMR, EXAFS/XANES, XPS, IR and UV-VIS spectroscopy, electron
microscopy, ESR, XRD, etc. Among these methods electronic spectroscopy plays
an important role. The UV-VIS spectra reflect changes in the oxidation state of
the metal as well as structural changes forced by incarceration and so serve as a
valuable tool for the ascertainment of intrazeolite complexation.Although vibra-
tional spectroscopy is most frequently applied, sometimes using the IR spectra
as fingerprints for identification, it is inadequate to predict the exact structure of
the clusters as these spectra may be different from those in solution or in the solid
state due to interaction with the zeolite matrix. In any case, reliable characteri-
zation requires the combined application of complementary analytical methods.
The preparation of small metal clusters with narrow particle size distributions
is of considerable interest from the catalytic point of view. One of the oldest
routes is the thermal expulsion of CO from a metal carbonyl. As the direct syn-
thesis is impractical for most of the d metals, the route often employed is reduc-
tive carbonylation, i.e., the reduction of a metal ion in the presence of CO. Ther-
mal treatment of Pt2+-exchanged faujasites, the properties of which were varied
from acidic to basic by neutralization of their protons by Na+ and Cs+ ions, in the
presence of CO at 373 K for several hours resulted in a red-purple coloring
accompanied by the appearance of two bands at 445 and 730 nm in the UV-VIS
spectrum in the case of the basic PtNaY sample [172]. Supported by IR spec-
troscopy and electron microscopy the carbonyl species formed were identified as
Chini complexes [173] of the general formula [Pt3(CO)3(2-CO)3]n2, structured
in a tinker toy construction, in line with n=3 and fitting into the supercage, while,
in the case of the other basic zeolite CsNaX, the entrapped species are suggested
to be extended through the 12R window in two adjacent supercages reaching
n=5. The formation of these anionic platinum carbonyl complexes in zeolite
PtNaX has been monitored in situ by combined electronic and IR spectroscopy
[174]. The monomer [Pt3(CO)3(2-CO)3]2 is distinguished by electronic bands at
318 and 456 nm remaining at their positions during reaction but growing with
time, as shown in Fig. 44. The additional absorption at 230 nm has been attrib-
uted to small Pt particles. The shoulder on the right flank of the 456-nm band
indicates the generation of oligomers with n1. These oligomers are more read-
ily formed in PtNaY, perceptible from additional bands at 384, 512 and 725 nm,
obviously due to the more acidic nature of this host.
Metal chelate complexes entrapped in zeolites have attracted increasing
attention for application as oxygen carriers, enzyme mimics and catalysts [48,
Fig. 44. DRS of reductive carbonylation of zeolite PtNaY at 363 K and p=105 Pa. In the lower
part the sequence of recording is 15 min during the first 4 h and thereafter at 8, 24 and 48 h in
the upper part. Reproduced from [174] by permission from Baltzer Science Publishers BV
175, 176]. Since 1977 several groups have succeeded in encapsulating tran-
sition metal phthalocyanines (Pc) in the supercages of zeolites A, X, Y and
VPI-5 using one of the synthetic strategies mentioned before. The zeolite
synthesis method, i.e., synthesis of the zeolite around FePc, CoPc and NiPc
complexes, has been applied by Balkus Jr et al. [177]. For these ship-in-a-bottle
species a strong Q band due to a pp* transition associated with the Pc
ring in the range 600900 nm is characteristic which, as shown in Fig. 45A
for FePc/NaX, appears as a broad absorption around 630 nm [178]. After
digestion of the sample in concentrated sulfuric acid the Q band near 790 nm
proves that the FePc is still intact (Fig. 45B). The frequency of this band is
sensitive to encapsulation and downscale shifted in comparison to the free
complex. In contrast to the interpretation of IR, XPS and EXAFS results, a
distortion from planarity of the MPc ligand could not be established by Raman
and NMR spectroscopy. Thus, it seems probable to rationalize this shift by
protonation of the MPc rather than by saddling as predicted by molecular
modeling [169].
Pd complexes of the Schiff base salen (N,N-bis-(salicylidene)-ethylenediamine)
and its chiral derivative salen* (R,R-N,N-bis-(3,5-di-tert-butylsalicylidene)-1,2-
cyclohexanediamine) have been immobilized in the large cages of zeolites FAU
and EMT, where the intracrystalline complex formation has been proved by UV-
VIS spectroscopy [179]. These host/guest compounds are active catalysts for
418 H. Frster
Fig. 45A, B. UV-VIS spectra of FePc inside zeolite NaX using A Nujol as mulling medium and
B FePc/NaX dissolved in sulfuric acid. Reprinted with permission from [178]. Copyright 1992
American Chemical Society
hydrogenations, the salen* complex being a potential candidate for enantiose-

lective hydrogenations.
A new family of hybrid organic-inorganic catalysts with dispersed active sites
inside ordered mesoporous materials has been prepared by anchorage of transi-
tion-metal ligands of Schiff base-type and chiral amino alcohols like (1R,2S)-
ephedrine on micelle-templated silicas [180]. Metalation of the grafted ligands
with manganese was followed by UV-VIS spectroscopy.
Microwave-assisted growth of 50-m large AlPO4-5 single crystals monomol-
ecularly loaded with dyes such as Rhodamine derivatives and Basic Yellow 40 has
been reported by Braun et al. [181]. UV-VIS reflectance as well as fluorescence
spectra revealed strong host/guest interaction resulting in distortion of the dye
molecule, presumably in antiparallel orientation to the polar axis of the host,
deviating from perfect alignment.While in the case of Rhodamine derivatives the
concentration-quenched emission is red shifted with increasing dye loading, this
phenomenon is reversed for Basic Yellow 40 uptake >4 mmol/g pointing to dis-
tortions of the host structure by the chromophore.
Molecular sieve encaged dye molecules are aspirants for photochemical appli-
cations in the fields of information processing and storage as well as for novel
pigments. As shown on the left-hand side of Fig. 46, e.g., thioindigo wrapped
up by the framework of zeolite NaX resides in the center of the 12R aperture
connecting two adjacent supercages, as revealed by Rietveld refinement of XRD
patterns [182]. The DRS of NaX-incarcerated thioindigo is characterized by the
strong S0S1 transition near 530560 nm (see Fig. 46, spectrum B) with a band-
width somewhat broader than in solution (Fig. 46, spectrum A), while adsorp-
tion at the external surface gives rise to further broadening together with a
bathochromic shift to 550 nm (Fig. 46, spectrum C) indicating stronger interac-
Fig. 46. Left Arrangement of trans-thioindigo between two supercages in zeolite NaX. The dots
mark the positions of the Na+ ions on SII* sites. Right UV-VIS spectra of thioindigo in toluene
solution (A), encaged in zeolite NaX (B) and adsorbed at the external surface (C). Reprinted
with permission from [182]. Copyright 1994 American Chemical Society
420 H. Frster
tion. The photoinduced trans-to-cis isomerization and the opposite thermal iso-
merization have been monitored by UV-VIS spectroscopy. The former is
assumed to start from the triplet state of the excited trans conformer and occurs
only provided the dye is not closely coordinated to the Na+ ions. Otherwise it is
prevented by strong interaction or by rapid quenching of the excited state. cis-to-
trans Isomerization, on the other hand, is retarded due to an enhanced energy
barrier for twisting.
In order to reduce the costs of photovoltaic devices for the conversion of light
to electricity a high-efficiency solar cell has been developed based on the photo-
sensitized electron injection into the empty states of nanosized semiconductors
using charge-transfer dyes as sensitizers [183]. Dye-sensitized cells consist of elec-
trodes based on an optically transparent film of titanium dioxide nanoparticles
coated with a monolayer of the respective dye. This assembly can be mimicked in
zeolites and related materials capable of hosting both semiconducting clusters and
charge-transfer dyes. Wark et al. [184] started their investigations applying the
mesoporous material MCM-41 as the host introducing Ti either by hydrothermal
synthesis or by impregnation. Dispersion and coordination of the Ti could be
inferred from the maxima and the onsets of absorption in the DR-UV/VIS spec-
tra, as has been shown in the preceding section. By hydrothermal synthesis most
of the titanium is tetrahedrally incorporated into the framework but restricted to
about 2 wt.%. Higher loadings, up to 8 wt.%, predominantly in octahedral coor-
dination are achieved by impregnation, where TiOx is bound to two silanol groups
at the same time.As the sensitizing dye zinc phthalocyanine and some of its deriv-
atives are encapsulated either admixed with the surfactant during MCM synthe-
sis or by impregnation. In the DR spectra the monomer M00 and M01 signals of
the Q band are observed representing the transitions from the HOMO to the
LUMO and its first overtone, pointing to the lack of aggregated ZnPc. Higher Ti
loading and a closer contact between the Ti center and the dye enable the charge
transfer from the latter to the former proved by the quenching of the fluorescence.
4.3
Concluding Remarks
Spectroscopic methods in general play an expanding role in characterization of

host/guest systems like, e.g., zeolite systems. They provide valuable microscopic
information on symmetry, structure and bonding in such associations, on the
location and motion of the guest, on the nature of active sites for catalytic trans-
formations, on the thus-formed intermediates and products, as well as on the
detection of poisons.
UV/VIS spectroscopy in particular is suited for the determination of the local
symmetry and the oxidation state of a TMI, but is not restricted to TMIs, and
gives information on the perturbation of the excited and ground electronic states
by host/guest interaction. There is a wide range of pressure and temperature in
which the experiments can be carried out without interference from the gas or
bulk solid phases, so that in situ as well as ex situ experiments are feasible.
UV/VIS spectroscopy has meanwhile become an established technique based
upon safe theoretical foundations. The experimental techniques are straight-
forward, the instrumental requirements modest. The spectra of even strongly

scattering samples can be readily interpreted and related to the properties of
these systems.
UV/VIS spectroscopy has proved to be a powerful tool for the detection of sta-
ble as well as reactive guests monitoring their change of concentration with time
of contact and temperature. By these means formation and further conversion of
intermediates can be pursued. However, from UV/VIS spectroscopy alone
detailed information on the species involved cannot be obtained. The combined
application of UV/VIS and other spectroscopic techniques (IR/Raman, NMR,
ESR) together with computer modeling and classical as well as quantum chem-
ical calculations raises hopes on yielding a deeper insight into the microscopic
world of host/guest systems.
5
References
1. Herzberg G (1966) Molecular spectra and molecular structure. III. Electronic spectra
and electronic structure of polyatomic molecules. Van Nostrand Reinhold Company,
New York
2. Jaff HH, Orchin M (1970) Theory and applications of ultraviolet spectroscopy. Wiley,
New York
3. Rao CNR (1975) Ultra-violet and visible spectroscopy, chemical applications. Butter-
worths, London
4. Lever ABP (1984) Inorganic electronic spectroscopy, 2nd edn. 2nd impression with
corrections 1986. Elsevier, Amsterdam
5. Suzuki H (1967) Electronic absorption spectra and geometry of organic molecules.
Academic Press, Inc., New York
6. Hollas JM (1987) Modern spectroscopy. Wiley, Chichester
7. Perkampus HH (1992) UV-VIS spectroscopy and its applications. Springer, Berlin
Heidelberg New York
8. Schlfer HL, Gliemann G (1967) Einfhrung in die Ligandenfeldtheorie. Akademische
Verlagsgesellschaft, Frankfurt am Main
9. Cotton FA (1971) Chemical applications of group theory. Wiley-Interscience, New York
10. Kober F (1979) Grundlagen der Komplexchemie. Salle+Sauerlnder, Frankfurt am Main
11. Uhlemann E (1977) Einfhrung in die Koordinationschemie. VEB Deutscher Verlag der
Wissenschaften, Berlin
12. Schuster P (1973) Ligandenfeldtheorie. Verlag Chemie, Weinheim
13. Kittel C (1973) Einfhrung in die Festkrperphysik. R. Oldenbourg, Mnchen
14. Leftin HP, Hobson MC Jr (1963) Adv Catal 14:115
15. Terenin A (1964) Adv Catal 15:227
16. Kiselev AV, Lygin VI (1975) Infrared spectra of surface compounds. Keter Publishing
House Ltd., Jerusalem
17. Chilton D, Rabinowitsch E (1932) Z Physik Chem B19:107
18. de Boer JH (1932) Z Physik Chem B18:49
19. de Boer JH, Custers JFH (1933) Z Physik Chem B21:208
20. Williams R (1989) Appl Spectrosc Rev 25:63
21. Threlfall TL (1988) Eur Spectros News 78:8 and references cited therein
22. Knowles A, Burgess C (eds) (1984) Practical absorption spectrometry. Chapman and Hall,
London
23. Dixit L, Ram S (1985) Appl Spectrosc Rev 21: 311
24. Talsky G (1994) Derivative spectrophotometry low and high order. VCH Verlags-
gesellschaft GmbH, Weinheim
422 H. Frster
25. Shibata S (1976) Angew Chem 88:750

26. Kortm G (1969) Reflexionsspektroskopie. Springer, Berlin Gttingen Heidelberg
New York
27. Sendoda Y, Ono Y, Keii T (1975) J Catal 39:357
28. Frster H, Schumann M, Seelemann R (1980) Stud Surf Sci Catal 5:211
29. Karge HG, aniecki M, Zioek M, Onyestyk G, Kiss A, Kleinschmit P, Siray M (1989) Stud
Surf Sci Catal 49:1327
30. Frster H, Franke S, Seebode J (1983) J Chem Soc Faraday Trans 1 79:373
31. Klier K (1967) Catal Rev 1: 207
32. Kellerman R (1979) Diffuse reflectance and photoacoustic spectroscopies. In: Delgass
WN, Haller GL, Kellerman R, Lunsford JH (eds) Spectroscopy in heterogeneous catalysis.
Academic Press, New York, p 86
33. Klier K (1980) Investigations of adsorption centers, molecules, surface complexes, and
interactions among catalyst components by diffuse reflectance spectroscopy. In: Bell AT,
Hair ML (eds) Vibrational spectroscopies for adsorbed species. ACS Symp Ser 137:141.
American Chemical Society, Washington
34. Schoonheydt RA (1984) Diffuse reflectance spectroscopy. In: Delannay F (ed) Charac-
terization of heterogeneous catalysts. Marcel Dekker, Inc., New York, p 125
35. Hurtubise RJ (1981) Solid surface luminescence analysis: theory, instrumentation,
applications. Marcel Dekker, New York
36. Vo-Dinh T (1984) Room temperature phosphorimetry for chemical analysis. Wiley,
New York
37. Svanberg S (1992) Atomic and molecular spectroscopy basic aspects and practical appli-
cations. Springer, Berlin Heidelberg New York London Paris Tokyo Hong Kong Barcelona
Budapest, p 329
38. Demas JN (1983) Excited state lifetime measurements. Academic, New York
39. Pott GT, Stork WHJ (1976) Catal Rev Sci Eng 12:163
40. Tanguay JF, Suib SL (1987) Catal Rev Sci Eng 29:1
41. Engel S, Kynast U, Unger KK, Schth F (1994) Stud Surf Sci Catal 84:477
42. Patarin J, Tulier MH, Durr J, Kessler H (1985) Zeolites 12:70
43. Leheny AR, Turro NJ, Drake JM (1992) J Chem Phys 97:3736
44. Beer R, Calzaferri G, Kamber I (1991) J Chem Soc Chem Commun 1489
45. Calzaferri G, Kunzmann A, Lain P, Pfanner K (1995) IS&Ts 48th Ann Conf Proc,
Washington DC 1995, p 318
46. Bedard RL, Bowes CL, Coombs N, Holmes AJ, Jiang T, Kirkby SJ, Macdonald PM, Malek
AM, Ozin GA, Petrov S, Plavac N, Ramik RA, Steele MR, Young D (1993) J Am Chem Soc
115:2300
47. Ozin GA, Kuperman A, Stein A (1989) Angew Chem 101:373
48. Ozin GA, Gil C (1989) Chem Rev 89:1749
49. Mortier WJ (1982) Compilation of extra framework sites in zeolites. Butterworth,
Guildford
50. Catlow CRA (ed) (1992) Modelling of structure and reactivity in zeolites.Academic Press,
London
51. Mikheikin ID, Zhidomirov GM, Kazansky VB (1972) Russ Chem Rev 41:468
52. Kellerman R, Klier K (1975) Surface and Defect Properties of Solids 4:1
53. Seff K (1976) Acc Chem Res 9:121
54. Lunsford JH (1976) Catal Rev 12:137
55. Lunsford JH (1977) ACS Symp Ser 40:473
56. Klier K, Hutta PJ, Kellerman R (1977) ACS Symp Ser 40:108
57. Atanasova VD, Shvets VA, Kazansky VB (1981) Russ Chem Rev 50:209
58. Schoonheydt RA (1993) Catal Rev Sci Eng 35:129
59. Sutton LE (1960) J Chem Educ 37:498
60. Carlin RL (1963) J Chem Educ 40:135
61. Verberckmoes AA, Weckhuysen BM, Schoonheydt RA (1997) Stud Surf Sci Catal
105:623
62. Uytterhoeven MG, Schoonheydt RA (1993) Probing the synthesis of CoAPO-n molecu-
lar sieves by analysis of wet precursor gels. In: von Ballmoos R, Higgins JB, Treacy MMJ
(eds) Proceedings from the Ninth International Zeolite Conference. Butterworth-Heine-
mann, Boston London Oxford Singapore Sydney Toronto Wellington, vol I, p 329
63. Han H-S, Chon H (1994) Stud Surf Sci Catal 84:797
64. Peeters MPJ, van Hooff JHC, Sheldon RA, Zholobenko VL, Kustov LM, Kazansky VB
(1993) Spectroscopic investigations of the redox properties of CoAPO molecular sieves.
In: von Ballmoos R, Higgins JB, Treacy MMJ (eds) Proceedings from the Ninth Interna-
tional Zeolite Conference. Butterworth-Heinemann, Boston London Oxford Singapore
Sydney Toronto Wellington, vol I, p 651
65. Gianotti E, Marchese L, Martra G, Coluccia S (1999) Reactivity of NO on Co2+/Co3+ redox
sites in CoAPO-18. FTIR and UV-Vis-NIR studies. In: Treacy MMJ, Marcus BK, Bisher ME,
Higgins JB (eds) Proceedings of the Twelfth International Zeolite Conference. Materials
Research Society, Warrendale, vol IV, p 2775
66. Echchahed B, Trong On D, Beland F, Bonneviot L (1999) Incorporation level and nature
of framework metal sites versus crystallization time and temperature in TS-1 and
FES-1. In: Treacy MMJ, Marcus BK, Bisher ME, Higgins JB (eds) Proceedings of the
Twelfth International Zeolite Conference. Materials Research Society,Warrendale, vol III,
p 1877
67. Reddy JS, Sayari A (1995) Stud Surf Sci Catal 94:309
68. Guo X, Wang X, Wang G, Li G (1997) Stud Surf Sci Catal 105:607
69. Hari Prasad Rao PR, Ueyama K, Kikuchi E, Matsukata M (1999) Synthesis of pure silica
beta and Al-free Ti-beta using TEAOH and their characterization. In: Treacy MMJ,
Marcus BK, Bisher ME, Higgins JB (eds) Proceedings of the Twelfth International Zeolite
Conference. Materials Research Society, Warrendale, vol III, p 1515
70. Corma A, Navarro MT, Prez-Pariente J, Snchez F (1994) Stud Surf Sci Catal 84:69
71. Franke O, Rathousk? J, Schulz-Ekloff G, Strek J, Zukal A (1994) Stud Surf Sci Catal 84:77
72. Corma A, Jorda JL, Navarro MT, Perez-Pariente J, Rey F (1999) Synthesis, characterization
and catalytic activity of Ti-MCM-41 materials obtained under highly acidic media. In:
Treacy MMJ, Marcus BK, Bisher ME, Higgins JB (eds) Proceedings of the Twelfth Inter-
national Zeolite Conference. Materials Research Society, Warrendale, vol II, p 817
73. Kumar R, Raj A, Kumar SB, Ratnasamy P (1994) Stud Surf Sci Catal 84:109
74. Bhaumik A, Kumar R, Ratnasamy P (1994) Stud Surf Sci Catal 84:1883
75. Kornatowski J, Wichterlova B, Rozwadowski M, Baur WH (1994) Stud Surf Sci Catal 84:
117
76. Concepcin P, Lpez Nieto JM, Prez-Pariente J (1995) Stud Surf Sci Catal 94:681
77. Patarin J, Tuilier MH, Durr J, Kessler H (1992) Zeolites 12:70
78. Weckhuysen BM, Schoonheydt RA (1994) Stud Surf Sci Catal 84:965
79. De Vos DE, Knops-Gerrits P, Parton RF, Weckhuysen BM, Jacobs PA, Schoonheydt RA
(1995) Coordination chemistry in zeolites. In: Herron N, Corbin DR (eds) Inclusion
chemistry with zeolites: nanoscale materials by design. Kluwer, Dordrecht, p 185
80. Packet D, Dehertogh W, Schoonheydt RA (1985) Stud Surf Sci Catal 24: 351
81. Kubelkova L, Jirka I, Vylita J, Novakova J (1994) Stud Surf Sci Catal 84:1051
82. Schnick W (1994) Stud Surf Sci Catal 84:2221
83. Klier K, Herman RG, Hou S (1994) Stud Surf Sci Catal 84:1507
84. Klier K (1971) Adv Chem Ser 101:480
85. Frster H, Thun K, Witten U (1983) Investigations on sorption and oxidation of carbon
monoxide in transition metal ion-exchanged zeolites A. Proceedings of the Fifth Inter-
national Symposium on Heterogeneous Catalysis, part 2, p 33
86. Frster H, Witten U (1987) Zeolites 7:517
87. Frster H, Schumann M, Seelemann R (1980) Stud Surf Sci Catal 5:211
88. Frster H, Seebode J (1983) Zeolites 3:63
89. Strome DH, Klier K (1980) ACS Symp Ser 135:155
90. Ebitani K, Morokuma M, Morikawa A (1994) Stud Surf Sci Catal 84:1501
91. Wichterlov B, Dedecek J, Sobalik Z (1995) Stud Surf Sci Catal 94:641
424 H. Frster
92. Borgmann C, Sauer J, Jstel T, Kynast U, Schth F (1999) The development of new lumi-
nescent materials from zeolite X. In: Treacy MMJ, Marcus BK, Bisher ME, Higgins JB (eds)
Proceedings of the Twelfth International Zeolite Conference. Materials Research Society,
Warrendale, vol III, p 2241
93. Naccache C, Chen FR, Coudurier G (1989) Stud Surf Sci Catal 49:661
94. Ramamurthy V, Eaton DF (1993) Dependence of the charge density and electric field
within the cages of X and Y zeolites on the cation: photophysical probes. In: von Ballmoos
R, Higgins JB, Treacy MMJ (eds) Proceedings from the Ninth International Zeolite Con-
ference. Butterworth-Heinemann, Boston London Oxford Singapore Sydney Toronto
Wellington, vol I, p 587
95. Ramamurthy V, Turro NJ (1995) Photochemistry of organic molecules within zeolites:
role of cations. In: Herron N, Corbin DR (eds) Inclusion chemistry with zeolites:
nanoscale materials by design. Kluwer, Dordrecht, p 239
96. Yoon KB, Kochi JK (1991) J Phys Chem 95:1348
97. Choi SY, Park YS, Yoon KB (1997) Stud Surf Sci Catal 105:671
98. Melsheimer J, Bhm MC, Lee JK, Schlgl R (1997) Ber Bunsenges Phys Chem 101:726
99. Melsheimer J, Schlgl R (1997) Ber Bunsenges Phys Chem 101:733
100. Karge HG, aniecki M, Zioek M (1988) J Catal 109:252
101. Karge HG, aniecki M, Zioek M (1986) Combined UV and IR spectroscopic studies on
the adsorption of SO2 onto faujasite-type zeolites. In: Iijima A,Ward JW (eds) New devel-
opments in zeolite science technology. Kondansha, Tokyo, p 617
102. Karge HG, Zioek M, aniecki M (1987) Zeolites 7:197
103. Abramov VN, Kiselev AV, Lygin VI (1963) Russ J Phys Chem 37:1507
104. Kupcha LA, Lygin VI, Mineeva LV (1968) Kinet Catal 9:691
105. Kiselev AV, Kupcha LA, Lygin VI, Shatskii VG (1969) Kinet Catal 10:370
106. Lygin VI (1971) Adv Chem Ser 102:86
107. Eremenko AM, Ogenko VM, Chuiko AA (1995) Stud Surf Sci Catal 94:607
108. Ignatovich M, Ogenko V, Chuiko A (1995) Stud Surf Sci Catal 94:615
109. Yoon KB (1993) Chem Rev 93:321
110. Olah G, Pittman CU Jr, Symons MCR (1968) Electronic spectra. In: Olah G, Schleyer P
(eds) Carbonium ions, vol I. Wiley-Interscience, New York, p 202
111. Karge HG (1973) Surf Sci 40:157
112. Garbowski ED, Praliaud H (1979) J Chim Phys Phys Chim Biol 76:687
113. Fejes P, Frster H, Kiricsi I, Seebode J (1984) Stud Surf Sci Catal 18:91
114. Frster H, Seebode J (1985) Spectroscopic and kinetic studies on the cyclopropane
isomerization over mordenites of different acidity. Proceedings of the International
Symposium on Zeolite Catalysis, Siofok. Acta Phys et Chem Szegediensis. p 413
115. Frster H, Seebode J, Fejes P, Kiricsi I (1987) J Chem Soc Faraday Trans 1 83:1109
116. Frster H, Kiricsi I, Seebode J (1987) Stud Surf Sci Catal 37:435.
117. Kiricsi I, Tasi Gy, Frster H, Fejes P (1990) J Mol Struct 218:369
118. Frster H, Kiricsi I (1987) Zeolites 7:508
119. Frster H, Kiricsi I (1988) Catalysis Today 3:65
120. Kiricsi I, Frster H (1988) J Chem Soc Faraday Trans 1 84:491
121. Kiricsi I, Tasi Gy, Fejes P, Frster H (1990) J Mol Catal 62:215
122. Kiricsi I, Frster H, Tasi Gy, Fejes P (1991) Stud Surf Sci Catal 65:697
123. Kiricsi I, Frster H, Tasi Gy, Fejes P (1989) J Catal 115:597
124. Kiricsi I, Tasi Gy, Frster H, Fejes P (1993) J Chem Soc Faraday Trans 89:4221
125. Kiricsi I, Frster H, Tasy Gy (1988) Stud Surf Sci Catal 46:355
126. Kiricsi I, Tasi Gy, Molnar A, Frster H (1990) J Mol Struct 239:185
127. Frster H, Kiricsi I, Tasi Gy, Hannus I (1993) J Mol Struct 296:61
128. Kiricsi I, Frster H, Tasi Gy (1991) J Mol Catal 65:L29
129. Vedrine JC, Dejaifve P, Garbowski ED, Derouane EG (1980) Stud Surf Sci Catal 5:29
130. Melsheimer J, Ziegler D (1992) J Chem Soc Faraday Trans 88:2101
131. Karge HG, Darmstadt H, Gutsze A, Vieth H-M, Buntkowsky G (1994) Stud Surf Sci Catal
84:1465
132. Kiricsi I, Frster H, Tasi Gy, Nagy JB (1999) Chem Rev 99:2085
133. Hannus I, Frster H, Tasi Gy, Kiricsi I, Molnar A (1995) J Mol Struct 348:345
134. Thomas JK (1993) Chem Rev 93:301
135. Enzel P, Bein T (1992) Synthesis of oligo- and polythiophenes in zeolite hosts. In: von
Ballmoos R, Higgins JB, Treacy MM (eds) Proceedings of the 9th International Zeolite
Conference. Butterworth-Heinemann, Boston London Oxford Singapore Sydney Toronto
Wellington, vol II, p 177
136. Bein T (1992) ACS Symp Ser 499:274
137. Cox SD, Stucky GD (1991) J Phys Chem 95:710
138. Stein A, Ozin GA (1993) Sodalite supralattices: from molecules to clusters to
expanded insulators, semiconductors and metals. In: von Ballmoos R, Higgins JB,
Treacy MMJ (eds) Proceedings from the Ninth International Zeolite Conference.
Butterworth-Heinemann, Boston London Oxford Singapore Sydney Toronto Wellington,
vol I, p 93
139. Herron N (1995) Zeolites as hosts for novel optical and electronic materials. In: Herron
N, Corbin DR (eds) Inclusion chemistry with zeolites: nanoscale materials by design.
Kluwer, Dordrecht, p 283
140. Sun T, Seff K (1994) Chem Rev 94:857
141. Beer R, Calzaferri G, Li J, Waldeck B. (1991) Coord Chem Rev 111:193
142. Bogdanchikova NE, Dulin MN, Toktarev AV, Shevnina GB, Kolomiichuk VN, Zaikovskii VI,
Petranovskii VP (1994) Stud Surf Sci Catal 84:1067
143. Linnert T, Mulvaney P, Henglein A, Weller H (1990) J Am Chem Soc 112:4657
144. Cvjeticanin ND, Petranovic NA (1994) Zeolites 14:35
145. Stucky GD, Srdanov VI, Harrison WTA, Gier TE, Keder NL, Moran KL, Haug K, Metiu HI
(1992) ACS Symp Ser 499:294
146. Srdanov VI, Blake NP, Markgraber D, Metiu H, Stucky GD (1994) Stud Surf Sci Catal
85:115
147. Blake NP, Stucky GD (1995) Alkali-metal clusters as prototypes for electron solvation in
zeolites. In: Herron N, Corbin DR (eds) Inclusion chemistry with zeolites: nanoscale
materials by design. Kluwer, Dordrecht, p 299
148. Edwards PP, Anderson PA, Thomas JM (1996) Acc Chem Res 29:23
149. Kasai PH (1965) J Chem Phys 43:3322
150. Liu X, Iu K-K, Thomas JK (1994) Chem Phys. Lett 224:31
151. Liu X, Iu K-K, Thomas JK (1994) J Phys Chem 98:13720
152. Liu X, Thomas JK (1992) Langmuir 8:1750
153. Iu K, Liu X, Thomas JK (1993) J Phys Chem 97:8165
154. Westphal U, Geismar G (1984) Z Anorg Allg Chem 508:165
155. Breuer REH, de Boer E, Geismar G (1989) Zeolites 9:336
156. Kodaira T, Nozue Y, Ohwashi S, Goto T, Terasaki O (1993) Phys Rev B 48:12245
157. Park YS, Lee YS, Yoon KB (1993) J Am Chem Soc 115:12220
158. Park YS, Lee YS, Yoon KB (1994) Stud Surf Sci Catal 84:901
159. Srdanov VI, Haug K, Metiu H, Stucky GD (1992) J Phys Chem 96:9039
160. Goto T, Nozue Y, Kodaira T (1993) Mat Sci Eng B 19:48
161. Brock M, Edwards C, Frster H, Schrder M (1994) Stud Surf Sci Catal 84:1515
162. Schrder M (1996) PhD thesis, University of Hamburg
163. Wang Y, Herron N (1988) J Phys Chem 92:4988
164. Herron N,Wang Y, Eddy MM, Stucky GD, Cox DE, Mller K, Bein T (1989) J Am Chem Soc
111:530
165. Klaas J, Kulawik K, Schulz-Ekloff G, Jaeger NI (1994) Stud Surf Sci Catal 84:2261
166. Grubert G, Wark M, Grnert W, Koch M, Schulz-Ekloff G (1997) Stud Surf Sci Catal
105:1077
167. Grubert G, Grnert W, Rathousky J, Zukal A, Schulz-Ekloff G, Wark M (1999) Structure
and redox properties of vanadium species in MCM-41. In: Treacy MMJ, Marcus BK, Bisher
ME, Higgins JB (eds) Proceedings of the Twelfth International Zeolite Conference.
Materials Research Society, Warrendale, vol II, p 825
426 H. Frster
168. Warnken M, Grubert G, Jaeger NI,Wark M (1999) Titanium and tin oxide-loaded zeolites
as optical sensor materials for reductive atmospheres. In: Treacy MMJ, Marcus BK, Bisher
ME, Higgins JB (eds) Proceedings of the Twelfth International Zeolite Conference.
Materials Research Society, Warrendale, vol III, p 2249
169. Balkus KJ Jr, Gabrielov AG (1995) Zeolite encapsulated metal complexes. In: Herron N,
Corbin DR (eds) Inclusion chemistry with zeolites: nanoscale materials by design. Kluwer,
Dordrecht, p 159
170. Kawi S, Gates BC (1994) Clusters in cages. In: Schmid G (ed) Clusters and colloids from
theory to applications. VCH, Weinheim, p 299
171. Ichikawa M (1992) Adv Catal 38:283
172. De Mallmann A, Barthomeuf D (1990) Catal Lett 5:293
173. Chini P (1980) J Organomet Chem 200:37
174. Schulz-Ekloff G, Lipski RJ, Jaeger NI, Hlstede P, Kubelkova L (1995) Catal Lett 30:65
175. Parton R, den Vos D, Jacobs PA (1992) Enzyme mimicking with zeolites. In: Derouane
et al. (eds) Zeolites; zeolite microporous solids: synthesis, structure, and reactivity.
Kluwer, Dordrecht, p 555
176. Herron N (1988) J Coord Chem 19:25
177. Balkus KJ Jr, Kowalak S, Ly KT, Hargis DC (1991) Stud Surf Sci Catal 69:93
178. Balkus KJ Jr, Hargis CD, Kowalak S (1992) ACS Symp Ser 499:347
179. Ernst S, Sauerbeck S, Yang X (1999) Preparation and catalytic properties of zeolite-
encapsulated palladium-salen-complexes in the hydrogenation of selected unsaturated
compounds. In: Treacy MMJ, Marcus BK, Bisher ME, Higgins JB (eds) Proceedings of the
Twelfth International Zeolite Conference. Materials Research Society,Warrendale, vol III,
p 2155
180. Sutra P, Bellocq N, Brunel D, Di Renzo F, Fajula F, Galarneau A, Lasperas M, Moreau P
(1999) Design of well-defined catalysts supported on micelle-templated silicas. In: Treacy
MMJ, Marcus BK, Bisher ME, Higgins JB (eds) Proceedings of the Twelfth International
Zeolite Conference. Materials Research Society, Warrendale, vol I, p 675
181. Braun I, Schomburg C, Bockstette M, Schulz-Ekloff G, Whrle D (1999) Novel pigments
via microwave-assisted crystallization inclusion of chromophores in AlPO4-5 or ship-in-
the-bottle synthesis of dyes in HY. In: Treacy MMJ, Marcus BK, Bisher ME, Higgins JB
(eds) Proceedings of the Twelfth International Zeolite Conference. Materials Research
Society, Warrendale, vol III, p 2233
182. Hoppe R, Schulz-Ekloff G, Whrle D, Kirschhock C, Fuess H (1994) Langmuir 10:1517
183. Hagfeldt A, Grtzel M (1995) Chem Rev 95:49
184. Wark M, Ortlam A, Ganschow M, Schulz-Ekloff G, Whrle D (1998) Ber Bunsenges Phys
Chem 102:1548

10 1007@b94239

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

10 1007@b94239

Uploaded by

Copyright:

Available Formats

Mol.

Sieves (2004) 4: 337 426

1 Introduction. Electronic Spectroscopy: Ranges, Limits,

2 Theoretical Fundamentals and Principles . . . . . . . . . . . . . . 339

3 Experimental Techniques . . . . . . . . . . . . . . . . . . . . . . . 365

Springer-Verlag Berlin Heidelberg 2004

4.1.2.1 General Aspects of the Influence of the Environment on the

In this chapter, ultraviolet(UV)-visible(VIS) spectroscopy of molecular sieves

where n is the frequency of the electromagnetic radiation, c is the speed of light

l = n lair and n c/cair = index of refraction (2)

Although in the UV-VIS region the radiation is more traditionally characterized

are also simultaneously excited, as illustrated in Fig. 1 by the electronic ground

A log (I0/I) = e C l (3)

The UV-VIS spectrum, usually an A or log A vs. n plot, in a first approximation

Bnm = (8 p3/4 p e0 3 h2) |Rnm|2 e (n) (4)

where e0 is the vacuum permittivity.

Rnm = y*n m ym dt (5)

The freedom of a possible transition is restricted by the selection rules, a kind of

Fig. 2. Term diagram for electronic transitions. A Absorption; F fluorescence; Ph phosphores-

states, so that the fluorescence transition is red-shifted compared to the excita-

Nv/Nv=0 = exp [(Ev Ev=0]/kT] (7)

higher vibrational levels v= 1, 2, are also appreciably populated, additional

Analogous to the concept of group vibrations in IR spectroscopy, photon absorp-

to lower or higher wavenumbers are called bathochromic or hypsochromic

Table 1. Different notations of electronic transitions

System Symbols Symbols Examples Correspondence

Kasha s, p, n s*, p* ss*

Table 2. The three lowest electronic transitions of ethylene in different notations

Configuration Kasha Mulliken Group theory Selection rule

Fig. 6. d Orbital splitting in ligand fields of different symmetries

In tetrahedral environments the splitting becomes reversed: the former t2g

with degeneracies given in parentheses.

In the strong-field limit, a d2 ion has the configurations:

(t2g)2 < (t2g)1 (eg)1 < (eg)2

A schematic overview of term splitting, starting from the strong-field limit

into consideration repulsion as well as field interaction, is given in Fig. 8.

the diagrams are in pairs qualitatively identical: those of the dn systems in an

b = B (complex)/B (free ion) < 1 (13)

The term b depends on the ligand descending in the so-called nephelauxetic

F > H2O > NH3 > CN, Cl > Br > I

Fig. 10. MO diagram of octahedral complexes under neglect of p electron interaction

Charge-transfer (CT) transitions take place in systems made up of two compo-

duration of collisions between donors and acceptors. It is suggested that donor-

From the viewpoint of the tight-binding approximation, the electronic structure

imperfections of structure or composition, normally present in all solids as a

The former is characterized by a radius small compared to the lattice para-

E =Eg 4 p2 e4/2 h2 e2 n2 (18)

In this section the influence of molecular interaction on solid surfaces in general

3. change of the bandshape with respect to its vibrational structure, and

The surfatochromic shift is characterized by magnitude and direction and is

hna h ng = QA QA* (19)

resulting in a frequency shift

Dn = na ng = (QA QA*)/h (20)

Generally, upon adsorption, the intensity of UV-VIS bands is significantly altered

Compared with IR spectroscopy the development of UV-VIS surface studies was

Spectroscopy of molecularly dispersed systems is routine both instrumentally

Since the triumphant advance of interferometers in the infrared range, spec-

Fig. 15. Schematic representation of the optical layout of a double-beam ratio-recording

In ordinary electronic spectroscopy the transmission mode is predominantly

Is = I0 24 p3 (n2 1/n2 + 2)2 NV2/l4 (21)

for the transmission experiments. Their general experience was a good

F(R) = (1R)2/2 R = K/S (23)

where R is the reflectance of an infinitely thick sample. The SKM function

Kasha s, p, n s, p ss*